Front cover

FICON Planning and Implementation Guide
Topologies, concepts, and terminology

Planning, implementation, and migration guidance Realistic examples and scenarios

Bill White Wolfgang Fries Brian Hatfield Michal Holenia Dennis Ng Ewerson Palacio René Petry

ibm.com/redbooks

International Technical Support Organization FICON Planning and Implementation Guide September 2009

SG24-6497-02

Note: Before using this information and the product it supports, read the information in “Notices” on page ix.

Third Edition (September 2009) This edition applies to FICON features defined as CHPID type FC, supporting native FICON, High Performance FICON for System z (zHPF), and FICON Channel-to-Channel (CTC) on IBM System z10 and System z9 servers.

© Copyright International Business Machines Corporation 2005, 2006, 2009. All rights reserved. Note to U.S. Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp.

Contents
Notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .x Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi The team who wrote this book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Become a published author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii Comments welcome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Part 1. Understanding FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Chapter 1. Introduction to FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1 Basic Fibre Channel terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.1.1 Node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1.2 Port . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1.3 Switched fabric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1.4 FC link . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1.5 World Wide Name . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 1.2 System z FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.2.1 FICON advantages over ESCON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.2.2 FICON operating modes and topologies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.2.3 Terms used with FICON Directors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.2.4 Terms used with the Input/Output architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Chapter 2. System z FICON technical description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Using the FICON architecture for I/O operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 FICON initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 FICON I/O request . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3 Command mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.4 Transport mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.5 Modified Indirect Data Address Word . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.6 Transport Indirect Data Address Word (TIDAW). . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.7 Open exchange. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.8 Buffer-to-buffer credit usage in FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.9 Extended distance FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 System z FICON feature support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chapter 3. FICON Director technical description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 The role of the FICON Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1 Switched point-to-point configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2 Cascaded FICON Director configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3 Basic components of a FICON Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.4 Basic functions of a FICON Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Qualified FICON Directors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 IBM System Storage SAN b-type family components. . . . . . . . . . . . . . . . . . . . . . 3.2.2 IBM System Storage SAN b-type family functions . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3 Cisco MDS 9500 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.4 Functions of the Cisco MDS 9500 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 14 14 16 19 20 22 22 23 23 24 25 33 34 34 35 36 37 44 45 47 52 53

Part 2. Planning the FICON environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

© Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

iii

Chapter 4. Planning the FICON environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Structured approach for planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.1 Migrating from ESCON to FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.2 Moving to a high bandwidth environment (FICON Express8) . . . . . . . . . . . . . . . . 4.4.3 Migrating from a single site to a multi-site environment . . . . . . . . . . . . . . . . . . . . 4.4.4 Implementing a new FICON environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5 Topologies and supported distances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.1 Point-to-point. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.2 Switched point-to-point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.3 Cascaded FICON Directors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.4 Extended distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6 Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.1 Intermix fabric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.2 Fabric security. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.3 High integrity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.4 Zoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7 Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7.1 Command-line interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7.2 Element management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7.3 Fabric management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7.4 Storage management initiative specification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7.5 System z management for FICON Directors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8 Virtualization and availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8.1 System z . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8.2 Control unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8.3 FICON Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9 Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.1 Frame pacing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.2 Extended distance FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.3 Multiple allegiance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.4 Parallel Access Volume and HyperPAV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.5 Modified Indirect Data Address Word . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.6 High Performance FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.7 Bandwidth management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.8 Traffic management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.9 Evaluation tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.10 Prerequisites and interoperability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.11 Physical connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

59 60 62 63 63 64 65 66 66 67 68 68 68 70 71 72 73 74 74 75 75 76 76 76 77 78 78 81 82 83 84 85 86 86 87 88 89 91 91 92 93

Part 3. Configuring the FICON environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Chapter 5. Configuring a point-to-point topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 5.1 Establishing a point-to-point topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 5.1.1 Description of our environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 5.1.2 Tasks and checklist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 5.1.3 Getting started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 Chapter 6. Configuring a switched point-to-point topology . . . . . . . . . . . . . . . . . . . . 6.1 Establishing a switched point-to-point topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1.1 Description of our environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1.2 Tasks and checklist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
FICON Planning and Implementation Guide

115 116 116 117

. . . . . . . . . . . 7. . . . . . . . . . . . . 7. . . . . . . . . . . . . . . . . . . . . . . . . 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 Tape . . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . .2. . . . . . . . 8. . . . . . . . . . . . .3. . . . . . Chapter 8. . . . . . 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . .9 Setting up the Allow/Prohibit Matrix (optional) . . . . . . . . . . . . . . . . . . . . 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. . . . .11 Configuring Port Fencing (optional) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 FICON Directors in an extended distance environment . . . . 199 Chapter 9. and DLS . . . .5. . . . . . . . . . . . . . . . 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8 Backing up Director configuration data . . . . . . . . . . . . . . . .4. . . . . 7. . . . . . . . . . . . . . . . . . . .1. . . . . . . . . . . . . . . Contents 201 202 203 203 204 205 205 206 206 207 209 210 211 212 213 213 v . . .2. . . . . . . .2 Using the DCFM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Setting up a logical switch (optional). . . . . . . . .5 ESS Link Statistics Report . . . .2 Resource Measurement Facility monitoring . . . . . . . . . . . . . . . . . . . . . 9. .2 Tasks and checklist. . . . . . . . . . . . . . 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . IOD.3. . . . . . . . . . . .7 Channel swap . . .5 RMF example reports . . . . . . . . . . . . . . . . . . . . . . . . .2 FICON Director management connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . .1. 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Configuration overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. . . . 8. . . . . . . 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . Managing the FICON environment . . . . . . . . . .3 Setting up a FICON Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. . . . . . . . . . . . . . . . .1 Configuration flowchart . . . .6 DCFM performance monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9. . .2 Enabling features (optional) . . .6 Enabling the Control Unit Port . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . . 8. . . . . . . . . . .3. . . . . . . . . . . 8. . . . . . . .1 Setting up Inter-Switch Links . . . . . . . . . . . 8. . . . . . . . . . . . . . . Monitoring the FICON environment . . . . . . . . . . . . . . . 8. .3. . . . . . . . . . . . .1. . . . . . Configuring a cascaded FICON Director topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 Configuring the Domain ID and Insistent Domain ID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. . . . .2 Installing and using Data Center Fabric Manager . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Data gathering . . .3. . . . 9. . 7. . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . 8. . . . . . . . 133 134 134 135 137 151 152 152 153 154 155 161 164 164 169 170 173 175 177 179 183 184 185 188 190 190 192 196 197 197 197 198 Part 4. . . .5 Setting up PBR.1 Description of our environment . .1 Installing Data Center Fabric Manager . .4 Setting up cascaded FICON Directors . . . . . . . . . . . .3 Introduction to performance monitoring . . .5. . . . . . 9. . . . . . . . . . . . . . . . . . . . . . . . . .8 Changing buffer credits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 General performance guidelines . . . . . .1 Establishing a cascaded FICON Director topology . .7 Changing port type and speed . . . . . . . . .1.5. . . .3. . . . . . . . 9. . . . . .1 DASD Activity Report . . . . . . . . . . . . . . . . . . . .2 I/O Queueing Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . .4 Introduction to Resource Measurement Facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. . . . . . . . . . . . . . . . . . . . . . . . .3 Getting started . . . . . . 119 Chapter 7. . . . . . 9. . . . .6. . . . . 9. . . Configuring FICON Directors . . . . . . . . . 9. .10 Setting up zoning (optional) . . . . . . . 8.4. . . . . . . . . . . . . . 8. . . . . . . . . . . . . . . . . . . . . . 8. . . . . . . . . . .6 FICON Directors in an intermixed environment. . . . . . . . . . . . 8. . . . . . . . . . . . . . . . . . . . .9 Backing up DCFM configuration data . . . . . . . . . . . 8. . . . . . . . . 9. . . . . . . . . 9. . . . . . . 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Getting started . . . . . . . . . . . . 8. . . .2 RMF reporting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1. . . . . 8. . . . . . . . . . . . . . . . . . .1.2 Setting up a high integrity fabric . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . .3 Channel Path Activity Report . . . . . . . . . . . . . . . . . 9. . .1 Changing the IP addresses. . . .4 FICON Director Report . . . . .1 System Activity Display . . . . . . . . . . . 8.

. . .1. . . . . . . . CHPID Mapping Tool . 10. . . . . . . . . . . . 10. . . . . . . . . . . . .2 Creating a CONFIG member . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 Appendix A. . . . . . . . . . . . . . . .1. . . . . . . . . . . . . . . . . 251 Appendix C. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DEVSERV command . . . . . . . . . . . . . . .9 DCFM data collection . . . . . . . . . . . . . . . I/O Configuration Program . . . . . . . . . . . . . . . . . Displaying the I/O configuration using D IOS. . . . . . . . . . . . .2 Real-Time Monitor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10. . . . . . . . . . . . . . . . . . . . . D M=DEV(dddd) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10. . . . . . . . . . . . .1 . . . . . . . . . . . . . . . . . . . . Displaying System z server CPC details. . . . . .1 Using the D M . . . . . . . . . . . . . . . . . . . . . 10. . .DS P. . . . . Displaying system status using D M=CPU .7 Node descriptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10. . . . . . . .interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Useful z/OS commands . . . . . . . . . . . Logging on to the HMC and SE . . . . . . . . . . . . . . . . . . . . Displaying the Analyze Channel Information panels. . . . 10. . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 Appendix E. . . . . . . . . .3 Environmental Record. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 265 266 266 266 268 270 272 287 288 288 290 290 291 292 292 296 300 vi FICON Planning and Implementation Guide . . . . . . . . . . . . . . . . . . . . . 243 Appendix B. Configuration worksheets . . . . . . . . . . . . . . . . . . . . Display units command D U . . . . . . . . . . . . . . . . . . . . . . Displaying individual FICON channel information . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . .dddd. . .8 DCFM logs . . . . . . . . . . . . . . IOCDS statements and keywords used for FICON. . . . . .Display system Matrix command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1. . . . . . . . . . . .2 Problem determination approach for FICON. . . .1 End-to-End Monitor . . . . . . . . . . . . . . . . . . . . . . Appendixes . and Printing program . . .9. . . . . . . . . . . . . . 259 Getting started . . . . . . . . . . Debugging FICON problems . . . . . . . . . Example: planning workflow . . . . .3 Most common I/O-related problems . . . . . . . . . . . . . . 217 218 218 218 219 220 222 222 222 225 226 226 228 229 230 Part 5. . . . Appendix F. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Using z/OS commands for problem determination. . . . . . . . . 214 Chapter 10. . . . . . . . Hardware Configuration Definition . . . . . 10. . . . . . . . . . .6 Helpful z/OS commands . . . . . . . . . . . . . . . . . . . . . . .CONFIG(HSA) . . . . . . . . . . . . . . . . . . . . 214 9. . . . . . .1 View node descriptors from DCFM . . . . . . . . . . . . . . . . . . . 10. . . . . . . . . . . . . . . . . . . . . 10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .CONFIG. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10. . . . Configuration and definition tools. . . . . . . . . . . . . . . . . . . 10. . . . . . . 255 256 256 256 256 257 Appendix D. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6. Hardware Configuration Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 Common z/OS FICON error message . . . . . . . Displaying additional z/OS information using D IPLINFO . . . . Configuring the DS8000 for FICON . . . . . . . . . . 10. . . . . . .4 FICON link incident reporting . . . . . . . . . . . . . . . . . . . . . . . . . Finding a physical resource on a System z server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Using HMC and SE for problem determination information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 FICON Purge Path Extended . . . . . .6.1 Preparing for problem determination activities . . . . . . . . . . .dddd. . . . . . . . .n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10. . . Editing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D U. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7. . . . Displaying HSA usage using D IOS. . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 DS8000 port layout . . . . . . . . HMC and SE versions and user IDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . .Appendix G. . . . . . . . . . . . Other publications . . . . . . . . . . . . . . Help from IBM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Online resources . . . . Adding FICON CTC connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 Contents vii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305 Related publications . . . . . . . . . . . . . . . . . . . . . . . How to get Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 307 307 309 309 309 Index . . . . . IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

viii FICON Planning and Implementation Guide .

IBM has not tested those products and cannot confirm the accuracy of performance. brands. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products. for the purposes of developing. IBM may have patents or pending patent applications covering subject matter described in this document. serviceability. program.A. All of these names are fictitious and any similarity to the names and addresses used by an actual business enterprise is entirely coincidental. or service that does not infringe any IBM intellectual property right may be used instead. therefore. 2009. this statement may not apply to you.S. EITHER EXPRESS OR IMPLIED. or service. NY 10504-1785 U. MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Any functionally equivalent product. You can send license inquiries. Information concerning non-IBM products was obtained from the suppliers of those products. in writing. program. THE IMPLIED WARRANTIES OF NON-INFRINGEMENT. IBM may use or distribute any of the information you supply in any way it believes appropriate without incurring any obligation to you. Consult your local IBM representative for information on the products and services currently available in your area. Changes are periodically made to the information herein. Any references in this information to non-IBM Web sites are provided for convenience only and do not in any manner serve as an endorsement of those Web sites. Some states do not allow disclaimer of express or implied warranties in certain transactions.A. these changes will be incorporated in new editions of the publication. All rights reserved. Any reference to an IBM product. which illustrate programming techniques on various operating platforms. program. IBM Corporation. The following paragraph does not apply to the United Kingdom or any other country where such provisions are inconsistent with local law: INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES THIS PUBLICATION "AS IS" WITHOUT WARRANTY OF ANY KIND. companies. The materials at those Web sites are not part of the materials for this IBM product and use of those Web sites is at your own risk. to: IBM Director of Licensing. using. or function of these programs. INCLUDING. You may copy. compatibility or any other claims related to non-IBM products. cannot guarantee or imply reliability. This information contains examples of data and reports used in daily business operations. IBM. IBM may make improvements and/or changes in the product(s) and/or the program(s) described in this publication at any time without notice.Notices This information was developed for products and services offered in the U. and distribute these sample programs in any form without payment to IBM. These examples have not been thoroughly tested under all conditions. program. and products. © Copyright IBM Corp. services. COPYRIGHT LICENSE: This information contains sample application programs in source language. modify. This information could include technical inaccuracies or typographical errors. Armonk. 2006. 2005. or service is not intended to state or imply that only that IBM product. or service may be used. it is the user's responsibility to evaluate and verify the operation of any non-IBM product. North Castle Drive. However. or features discussed in this document in other countries. the examples include the names of individuals. IBM may not offer the products. To illustrate them as completely as possible.S. BUT NOT LIMITED TO. their published announcements or other publicly available sources. therefore. The furnishing of this document does not give you any license to these patents. marketing or distributing application programs conforming to the application programming interface for the operating platform for which the sample programs are written. ix .

and the NetApp logo are trademarks or registered trademarks of NetApp.com are trademarks or registered trademarks of International Business Machines Corporation in the United States. and the IntelliMagic logo are trademarks of IntelliMagic BV in the United States. Java. Inc. in the United States. Convergence. other countries. SAP. and ibm. Active Directory. Microsoft. Such trademarks may also be registered or common law trademarks in other countries. Other company. Linux is a trademark of Linus Torvalds in the United States. other countries. other countries.com/legal/copytrade. or both. These and other IBM trademarked terms are marked on their first occurrence in this information with the appropriate symbol (® or ™).Trademarks IBM. other countries. or service names may be trademarks or service marks of others. or both. or both. and all Java-based trademarks are trademarks of Sun Microsystems. A current list of IBM trademarks is available on the Web at http://www. other countries. product. NOW. other countries. or both. or both. the IBM logo. and other countries. Excel. and the Windows logo are trademarks of Microsoft Corporation in the United States. or both: DB2® DS8000® ESCON® FICON® GDPS® Geographically Dispersed Parallel Sysplex™ HyperSwap® IBM® Parallel Sysplex® PR/SM™ Redbooks® Redbooks (logo) Resource Link™ System Storage™ System z10™ System z9® System z® Tivoli® z/Architecture® z/OS® z/VM® z/VSE™ z9® ® The following terms are trademarks of other companies: Disk Magic. in the U. Snapshot.shtml The following terms are trademarks of the International Business Machines Corporation in the United States.ibm. and SAP logos are trademarks or registered trademarks of SAP AG in Germany and in several other countries. indicating US registered or common law trademarks owned by IBM at the time this information was published. Inc. x FICON Planning and Implementation Guide .S.

France. implementation. He has 11 years of field experience with IBM. and FICON Directors. Poughkeepsie Center. MAN and Private Optical Network WDM solutions. Dennis Ng is a Consulting IT Specialist with IBM North America. Bill White is a Project Leader and Senior Networking and Connectivity Architect at the International Technical Support Organization. FICON Channel-to-Channel (FCTC). and Certified Information Systems Security Professional. data center planners.You will find configuration examples required to support FICON control units. His areas of expertise include Systems z hardware and FICON/ESCON® connectivity technical support. It discusses the FICON and Fibre Channel architectures. xi . All rights reserved. The team who wrote this book This book was produced by a team of specialists from around the world working at the International Technical Support Organization. including working with Networking and WDM products. The book focuses on the hardware installation and the software definitions that are needed to provide connectivity to FICON environments. 2009. and he currently develops and delivers technical training for the System z environment. ADVA Certified Engineer. Maryland. 2005. SAN. Dennis is currently working at the Washington Systems Center in Gaithersburg. implementation. He is a Cisco Certified Internetwork Expert. Michal Holenia is a Senior Network IT Architect and Integrated Communication Services Leader in IBM Slovakia. as well as a broad understanding of the Fibre Channel and FICON architectures. Poughkeepsie Center. and technical support for LAN. He has 34 years of IBM experience in Large Systems Technical Support. Wolfgang has 32 years of experience in supporting large System z customers. HCD or IOCP. It also discusses utilities and commands that are useful for monitoring and managing the FICON environment. 2006. The target audience for this document includes IT Architects. Ewerson Palacio is an IBM Senior Technical Staff Member (STSM) and a Certified Consulting IT Specialist for Large Systems in Brazil. terminology. SAN administrators. and supported topologies. Wolfgang Fries is a Senior Consultant in the System z Support Center in Germany. His areas of expertise include System z servers and connectivity. He has more than 30 years of experience in the IBM mainframe environment and began his career as a Large System Customer Engineer in Southern California. Brian has been in the Education field for the past 17 years. focusing on FICON architecture and FICON supported products. Michal provides design. WAN. and system programmers who plan for and configure FICON environments.Preface This IBM® Redbooks® publication covers the planning. and management of IBM System z® FICON® environments. Georgia. He has more than 35 years of © Copyright IBM Corp. The reader is expected to have a basic understanding of IBM System z10™ and IBM System z9® hardware. He spent several years in the European Support Center in Montpellier. providing international hardware support for System z servers. Brian Hatfield is a Certified Consulting Learning Specialist who works for the IBM Systems and Technology Group in Atlanta.

FICON Implementation Guide. Thanks to the following people for their contributions to this project: Bob Haimowitz International Technical Support Organization. You will have the opportunity to team with IBM technical professionals. Poughkeepsie Jack Consoli Systems Engineer. were: Hans-Peter Eckam. Poughkeepsie Charlie Hubert. browse the residency index. IBM Germany Iain Neville. Fibre Channel Directors. published in February 2005.com/redbooks/residencies. Brian Jacobs. developer. Since serving a three-year apprenticeship at IBM in Mainz. FICON Implementation Guide. were: Hans-Peter Eckam. IBM UK Authors of the second edition. Poughkeepsie Center Connie Beuselinck IBM System z Product Planning. Find out more about the residency program. Thanks to the authors of the previous editions of this book. you will develop a network of contacts in IBM development labs. Brocade Communications Systems. and Clients. and increase your productivity and marketability. Your efforts will help increase product acceptance and customer satisfaction. and instructor for the last five generations of IBM high-end servers. Rene has worked for nine years as an Account Customer Engineer for a large banking customer in Germany. As a bonus. and apply online at: ibm.experience with IBM Large Systems. Ewerson has been a System z hardware Top Gun course designer. His areas of expertise include System z servers. Inc. IBM Germany Wolfgang Fries. while getting hands-on experience with leading-edge technologies. FICON. Sam Mercier IBM Vendor Solutions Connectivity (VSC) Lab.to six-week residency program! Help write a book dealing with specific products or solutions. His areas of expertise include System z server technical and customer support. Authors of the first edition. IBM UK Become a published author Join us for a two. Poughkeepsie Lou Ricci IBM Systems Software Development. IBM Storage products (disk and tape).html xii FICON Planning and Implementation Guide . Business Partners. and fiber optic infrastructures. published in January 2006. IBM Germany Iain Neville. Rene Petry is a Support Center Specialist in the System z Support Center in Germany.

ibm. Send us your comments about this book or other IBM Redbooks publications in one of the following ways: Use the online Contact us review Redbooks form found at: ibm. HYTD Mail Station P099 2455 South Road Poughkeepsie.Comments welcome Your comments are important to us! We want our books to be as helpful as possible. International Technical Support Organization Dept.com Mail your comments to: IBM Corporation.com/redbooks Send your comments in an e-mail to: redbooks@us. NY 12601-5400 Preface xiii .

xiv FICON Planning and Implementation Guide .

1 . 2006. It also provides information about the FICON and Fibre Channel architectures and their uses in System z environments. the FICON Director. © Copyright IBM Corp.Part 1 Part 1 Understanding FICON This part introduces FICON and explains how it is exploited by the System z channel. and the control unit. All rights reserved. 2009. 2005.

2 FICON Planning and Implementation Guide .

FICON Express4. 2005. Throughout this chapter we use the term FICON to refer to FICON Express8.1 Chapter 1. 2009. This chapter discusses the basic Fibre Channel (FC) and FICON terminology. as well as System z FICON support. All rights reserved. and FICON Express. and topologies. 2006. and published as ANSI standards. operating modes. Introduction to FICON The term Fibre Connection (FICON) represents the architecture as defined by the InterNational Committee of Information Technology Standards (INCITS). except when the function being described is applicable to a specific FICON feature type. 3 . FICON Express2. FICON also represents the names of the various System z server I/O features. © Copyright IBM Corp. benefits.

Encoding is done by the N_Port when sending the character stream over the fiber. as illustrated in Figure 1-1.Interface and Media The Fibre Channel physical interface (FC-0). specified in FC-PI.1 Basic Fibre Channel terminology This section describes the Fibre Channel (FC) architecture and explains general terms that are used both in the FC and FICON environments. and the decoding back to 8-bit code is performed by the receiving N_Port. and long-implemented channel and networking protocols to run over the same physical interface and media. consists of the transmission media. and peripherals developed by the InterNational Committee for Information Technology Standards (INCITS) and American National Standard Institute (ANSI). receivers. well-known. 4 FICON Planning and Implementation Guide .2 Networks FC-4 Mapping Multimedia FC-3 FC-2 Signaling Protocol Common Services Framing Protocol/Flow Control FC-1 Transmission Protocol Encode/Decode (8/10) FC-0 Interface/Media Physical Characteristics Copper/Multi-Mode Fiber/Single-mode Fiber Figure 1-1 Fibre Channel architectural levels The FICON channel architecture consists of the following Fibre Channel (FC) protocols: FC-0 level . The physical interface specifies a variety of media and associated drivers and receivers capable of operating at various speeds. there is a rich set of standards and terminology associated with its architecture.1. devices. is an integrated set of rules consisting of five layers (FC-0 through FC-4) for serial data transfer between computers. FC-1 level . Audio / Video FC-FCP FICON IPI / SCSI / HIPPI / SB / IP / SB-3/4 Channels Upper Level Protocol (ULP) P / 802. and their interfaces. a unique bit-pattern is assigned to each known hexadecimal character.Transmission Protocol This is a link control protocol that performs a conversion from the 8-bit EBCDIC code into a 10-bit transmission code. One defining characteristic of Fibre Channel is that it allows many existing. Consequently. transmitters. FICON uses the Single-Byte Command Code Sets -3 (SB-3) and Single-Byte Command Code Sets -4 (SB-4) implementation within the Fibre Channel standard. The FC architecture.

– FICON was introduced with Single Byte-2 Command Code Sets mapping protocol. FC-4 level . as shown in Figure 1-2. and FC-SB-4 architecture information and other FC documentation can be obtained from the following Web site: http://www. FC-SB-3. FC-SB-2/3/4 (FICON). In 2008 another revision was drafted.Signaling Protocol Fibre Channel physical framing and signaling interface (FC-PH) describes the point-to-point physical interface. transmission protocol. identified as FC-SB-4. that can operate in a point-to-point. and then later revised with FC-SB-3.t11. Chapter 1. Introduction to FICON 5 . and others. and signaling protocol of high-performance serial links for support of higher-level protocols associated with HIPPI. arbitrated loop. FC-SB-2. or switched configuration.Common Services This layer is intended to provide the common services required for advanced features.org Fibre Channel provides the capability to build a configuration.FC-2 level . Switched Fabric FC Link F_Port E_Port E_Port N_Port F_Port N_Port ISL F_Port FC Link N_Port FL_Port NL_Port NL_Port NL_Port NL_Port Figure 1-2 FC configuration with port types Note: System z FICON channels do not support the arbitrated loop topology.Mapping The Upper Level Protocol (ULP) is part of FC-4 and describes IPI/FC-FCP (SCSI)/HIPPI/SB/IP and FICON. SCSI. this revision will be used to support additional FICON functions. IPI. FC-3 level .

a device controller. or a peripheral device (such as a disk or tape drive).1. Switches may be interconnected with a Inter-Switch Link (ISL) to which the E_Ports are connected. 1. Other port types include: E_Port F_Port FL_Port G_Port L_Port NL_Port U_Port An expansion port is used to interconnect switches and build a switched fabric. 6 FICON Planning and Implementation Guide .1.1. The port type is determined by the node’s role in the topology. 1. an F_Port. 1.2 Port Each node must have at least one port (hardware interface) to connect the node to the FC topology. as shown in Figure 1-2 on page 5. This node port is referred to as an N_Port. The FC link is a fiber optic cable that has two strands.1.3 Switched fabric One or more switches are interconnected to create a fabric. One strand is used to transmit a signal and the other strand is used to receive a signal (see Figure 1-3 on page 7). A node has a unique 64-bit identifier known as the Node_Name. It is a generic switch port that can operate as an E_Port. which is a unique 64-bit identifier that is assigned at the time it is manufactured. The N_Port is used to associate an access point to a node’s resources. A node loop port is an N_Port with loop capabilities. A switched fabric takes advantage of aggregated bandwidth via switched connections between N_Ports. It can be a computer (host).4 FC link The port connects to the topology through an FC link.1. A generic port is a port that has not yet participated in a fabric. or an FL_Port. A universal port is a port that has not yet assumed a specific function in the fabric. A fabric port is used to connect an N_Port to a switch that is not loop-capable.1 Node A node is an endpoint that contains information. Each N_Port has a Port_Name. A loop port is a port in a Fibre Channel Arbitrated Loop (FC-AL) topology. An FC link is used to interconnect nodes and switches. The Node_Name is typically used for management purposes. A fabric loop port is used to connect NL_Ports to a switch in a loop configuration. to which the N_Ports are connected.

and when they are worldwide unique. These addresses are assigned by the manufacturer.5 World Wide Name Nodes and ports have unique 64-bit addresses that are used to identify them in an FC topology.FC port FC port Outbound Outbound Tx Rx Tx Rx Inbound Inbound Fibre Channel link Figure 1-3 Fibre Channel link For example. Introduction to FICON 7 . Figure 1-4 illustrates where the WWNs are used. separated by colons (for example. 08:45:12:56:43:00:D5:A0). they are referred to as: World Wide Node_Name (WWNN) World Wide Port_Name (WWPN) A WWN (any WWNN or WWPN) is usually written in sets of two hex digits. an FC link (port-to-port connection) can be: Node-to-node link (N_Port-to-N_Port) Node-to-switch link (N_Port-to-F_Port) Switch-to-switch link (E_Port-to-E_Port) 1.1. Server (Node) WWNN1 WWPN11 WWPN31 Controller (Node) WWNN3 N_Port WWPN21 Switch (Node) WWNN2 WWPN22 N_Port F_Port F_Port F_Port WWPN23 F_Port WWPN24 N_Port WWPN12 N_Port WWPN32 Figure 1-4 World Wide Names Chapter 1. These addresses (in the FC standard) are called Node_Names and Port_Names. with a vendor-specific portion defined by the IEEE standards committee.

The I/O adapter may support various FICON operating modes. HyperPAVs can be used to improve I/O efficiency and performance. 1. Other FICON enhancements may require the use of certain FICON features used on a specific System z server. Refer to IBM System Storage DS8000: Architecture and Implementation. such as the System z10 or System z9 servers. In this book we focus on the FICON native operating mode and the latest FICON capabilities available on the System z10 and System z9 servers. FICON is implemented differently from other FC implementations.On a System z server. This approach is used by all Fibre Channel protocol Host Bus Adapters (HBAs) users connected to a switched point-to-point topology Definition Uses the Hardware Configuration Definition (HCD) tool or Input Output Configuration Program (IOCP) to define the associated N_Ports to which communication is allowed. while simultaneously reducing the number of I/O resources used in System z environments. The definitions in HCD/IOCP must match the actual port connections of the attached control units. and provides additional strengths and capabilities compared to the ESCON technology. FICON is widely used in the System z environment. the WWNN is constant for all FICON channels (ports). Discovery Uses the WWN and discovers the associated N_Port address by querying the Fabric Name Server. Some control units and control unit functions may require FICON use exclusively. Hyper Parallel Access Volume requires the use of FICON and will not work with ESCON. Additional support or licenses may be required on the control unit. in a switched point-to-point topology. however. For example. and servers. the WWPN is unique to each FICON channel on the server. many of the FC terms and definitions are used in the various FICON configurations. Hyper Parallel Access Volume (HyperPAV) Hyper Parallel Access Volume (HyperPAV) is an enhancement to the original PAV feature supported on DASD subsystems.2 System z FICON Because FICON is an IBM I/O technology built on the Fibre Channel (FC) architecture. Many additional capabilities have been included in support of FICON since it was originally introduced. switches. System z FICON only uses the definition process. HyperPAVs require FICON connections. I/O adapters or ports supporting FICON reside on control units. For example. FICON channel architecture is compatible with: Fibre Channel Physical and Signaling standard (FC-FS) Fibre Channel Switch Fabric and Switch Control Requirements (FC-SW) Fibre Channel Single-Byte-3 (FC-SB-3) and Fibre Channel Single-Byte-4 (FC-SB-4) Even though FICON uses the FC architecture to transport commands and data over the fiber link. SG24-6786 for more information about HyperPAV. 8 FICON Planning and Implementation Guide . a server using FC may use a discovery or definition process to obtain a port address.

FICON-to-ESCON solutions The FICON Express feature was the last FICON channel feature to support CHPID type FCV. which is a significant enhancement compared with ESCON channels. and multiplexing of mixed workloads. Table 1-1 FICON-to-ESCON comparison ESCON Channel concurrent I/O operations Command execution Fiber use Link data rate Switching Unrepeated distance from channel Max distance without performance degradation Unit addresses per channel One Synchronous Channel/CU handshake Half duplex 200 Mbps Circuit switching 3 km 9 km 1024 FICON Up to 32 or more depending on FICON feature and servera Asynchronous Full duplex Up to 8 Gbps b Packet switching 10 km c 100 km or more d 16384 a. c. FICON Express8 provides up to 8 Gbps link rate. increased bandwidth. The FICON Express features are supported on System z10 and System z9 servers only if carried forward on a server upgrade. The data center I/O configuration now has increased connectivity flexibility because of the increased I/O rate. Many System z environments only use FICON Express4 features and above. and more addresses as shown in Table 1-1. Chapter 1. FICON channels also permit intermixing of large and small I/O operations on the same link. This is one of the fundamental differences between FICON and ESCON. FICON switches.2. Standard unrepeated distance for most FICON LX features is 10 km. Therefore. The PRIZM FICON to ESCON Converter is a channel-based appliance that converts native FICON protocol to native ESCON protocol. It enables multiple concurrent I/O operations to occur simultaneously to multiple control units. thereby allowing attachment of existing ESCON devices directly to FICON channels. To avoid performance degradation at extended distances. Maximum number of concurrent I/O operations can vary depending on the use of high performance FICON and FICON feature. The maximum link rate depends on the specific FICON feature.1. additional distance may be achieved with an RPQ on some FICON features. such as the PRIZM FICON to ESCON Converter from Optica Technologies Inc. thereby enabling improved performance. The FICON Express channel in FCV mode allows the accessing of ESCON interface control units with ESCON interfaces via a 9032-5 ESCON Director FICON Bridge adapter. FICON uses fiber cabling more efficiently. longer distances. d. with its sequential I/O operations. another solution is needed for FICON channel-to-ESCON control unit connectivity. Introduction to FICON 9 . directors. b. or channel extension devices may be required.1 FICON advantages over ESCON FICON offers many advantages over ESCON.

The three operating modes that can be defined are: FICON conversion mode (FCV) – This operating mode is only supported on the FICON Express feature FICON native mode (FC) Fibre Channel Protocol (FCP) As shown in Figure 1-5. PRIZM supports native FICON channel protocol (CHPID type FC). The defined operating mode supported depends on the FICON feature installed in the System z server. and supports attachment to existing FICON fabrics.2. For additional details.wss/itservice/igs/a1026000 1. PRIZM provides strategic flexibility to those migrating from ESCON to FICON on their System z servers. refer to the following Web site: http://www.ibm.2 FICON operating modes and topologies System z FICON features may be defined in HCD or IOCP to operate in one of up to three operating modes.com/?page_id=19 The PRIZM solution is also offered through IBM GTS Site and Facilities Services: http://www.Unlike the Model 5 ESCON Director FICON Bridge solution. a FICON channel in native mode can access FICON control units in any of the following three topologies: Point-to-point (direct connection) Switched point-to-point (via a FICON Director) Cascaded FICON Directors (through two FICON Directors) Point-to-point FC FC Link FICON CU Switched point-to-point FC FC Link FC Link FICON CU FICON CU FC L ink FICON Director FC Lin k FC Cascaded FICON Directors Site A Site B ISL F FC FC Lin k k C Lin FICON CU FICON Director FICON Director FC L in k nk FICON CU FC Figure 1-5 System z supported FICON native topologies 10 FICON Planning and Implementation Guide FC Li .opticatech.com/services/us/index.

Introduction to FICON 11 . They provide a pipeline through which data is exchanged between servers. can also communicate with other FICON control units. or between a server and external devices. The switch control unit port (CUP) function allows z/OS® to manage a FICON Director with the same level of control and security as for an ESCON switch. an I/O device is attached to one control unit and is accessible through one channel path. It provides the data path and synchronization for data transfer between two channels. Switch ID and switch address (1-byte value) are terms used to address a FICON Director. z/Architecture channel connections are referred to as channel paths. The FICON channel at each end of the FICON CTC connection. 1. Host communication includes control functions such as blocking and unblocking ports. When the CTC option is used to connect two channels that are associated with different systems. Input/output channels Input/output (I/O) channels are components of the System z Channel Subsystem (CSS) and z/Architecture®. We define some of those terms here. The entry switch and cascaded switch are interconnected via an Inter-Switch Link (ISL). 1.3 Terms used with FICON Directors It is important to point out that some definition terms used in IOCP and HCD were carried over from the ESCON environment.2. a loosely coupled Chapter 1. or a means of communication between a system and its environment. Control unit The most common attachment to a System z channel is a control unit (CU) accessed via an Enterprise Systems CONnection (ESCON) or a FIbre CONnection (FICON) channel. supporting the FCTC control units. we use the following terms throughout this book to describe the definitions needed in IOCP and HCD for cascaded FICON Directors and switched point-to-point configurations: An entry switch is the FICON Director that is directly connected to the processor’s FICON channel and to the CU (destination) or another FICON Director. In the simplest case. A cascaded switch is the FICON Director that connects to the CU (destination) and to the entry switch. A port address (1-byte value) is used to address the physical port on the FICON Director.4 Terms used with the Input/Output architecture The System z channel environment uses many terms and acronyms. I/O devices An input/output (I/O) device provides external storage. Channel-to-Channel The Channel-to-Channel (CTC) function simulates an I/O device that can be used by one system control program to communicate with another system control program. The terms listed here will be used throughout this book as we discuss channel and I/O operations in the various chapters. The CU controls I/O devices such as disk and tape devices. such as disk and tape devices. as well as monitoring and error reporting functions. For clarity.A FICON channel also supports channel-to-channel (CTC) communications.2. a means of communication between data processing systems.

as viewed by either of the channels it connects. Multiple CSSs can be configured within the same System z (up to four on z10 EC servers and z9 EC servers. 12 FICON Planning and Implementation Guide . Channel Path Identifier A Channel Path Identifier (CHPID) is a value assigned to each channel path of the system that uniquely identifies that path. Two subchannel sets are provided: subchannel set-0 can have up to 63. The CSS has evolved with the increased scalability of IBM System z servers. A subchannel is assigned for each device defined to the logical partition. Channel spanning Channel spanning extends the MIF concept of sharing channels across logical partitions to sharing channels across Channel Subsystems and logical partitions. and subchannel set-1 can have up to 64 K subchannels. has the appearance of an unshared input/output device. A total of 256 CHPIDs are supported by the CSS.multiprocessing system is established. and up to two on z10 BC servers and z9 BC servers). Subchannels A subchannel provides the logical representation of a device to the program. Multiple Image Facility Multiple Image Facility (MIF) enables resource sharing across logical partitions within a single CSS or across CSSs. The CTC connection. Channel Subsystem The Channel Subsystem provides the functionality for System z servers to communicate with input/output (I/O) devices and the network.75 K subchannels. The CSS architecture provides functionality in the form of multiple Channel Subsystems (CSSs). Note that Multiple Subchannel Sets (MSS) are available on System z10 and System z9 servers to increase addressability. It contains the information required for sustaining a single I/O operation.

FC-SB-4 is intended to be a complete replacement of the FC-SB-3 standard.2 Chapter 2. System z FICON technical description FICON uses the Fibre Channel Single Byte Command Sets-3 (FC-SB-3) standard as defined by the InterNational Committee of Information Technology Standards (INCITS). This chapter provides a technical description of how System z servers use the FICON architecture to process I/O operations. All rights reserved. 2009. 2005. as well as the FICON features and functions available on System z10 and System z9 servers. © Copyright IBM Corp. 2006. zHPF uses the architecture as described in the draft proposal for Fibre Channel Single Byte Command Sets-4 (FC-SB-4 ). Late in 2008 IBM introduced System z High Performance FICON (zHPF) on the System z10. and published as ANSI standards. 13 .

when supported by the channel and control unit Channel logical-path establishment For a control unit. the initialization process for a FICON channel consists of the following procedures: Link initialization Channel login and security attribute determination Channel node-identifier acquisition Channel state-change registration Channel link-incident-record registration Process login. and is not part of the execution of an I/O operation. If a switch fabric is present. the initialization process occurs infrequently. Both modes (command and transport) are discussed in greater detail later in this chapter. If a switch fabric is not present. the link must be initialized. exploiting the FC-SB-4 protocols.1 FICON initialization The FICON channel initialization process establishes the necessary conditions for elements of a channel path to be able to sustain both link-level and device-level communication. the System z10 FICON channel will operate in command mode. the channel will perform the I/O operation in transport mode. If the control unit does not support zHPF. Note the following points: When both the System z10 FICON channel and the control unit indicate support for zHPF. it is usually required only after a power-on procedure or during a system initialization procedure. the channel performs an N_Port Login.2. For example. For example.1. if supported Control unit logical-path establishment The Fabric login (FLOGI) and Port login (PLOGI) process will determine the presence of a switch fabric. After the initialization process is complete. 14 FICON Planning and Implementation Guide .1 Using the FICON architecture for I/O operations Before a System z FICON channel can send any I/O over the link. There are specific steps used in this initialization process. There is a set of initialization steps that is used with every connection. Under normal conditions. zHPF must be enabled in z/OS to perform transport mode operations. A System z10 FICON channel can operate in command mode and in transport mode (also known as zHPF) simultaneously. the initialization process consists of the following procedures: Link initialization Control unit login Control unit node-identifier acquisition Control unit state-change registration Process login. the channel performs an F_Port Login. and they are described in detail in the next section. There is also a series of steps used for every control unit image stored in the HSA. I/O operations can be started on the System z FICON channel. exploiting the FC-SB-3 protocols. A FICON channel operates in what is known as command mode. 2.

Figure 2-1 shows the key steps used in the FICON initialization process for a switched-point-to-point connection with a control unit that supports zHPF. 6. ELP to the control unit images that are stored in the HSA. Switched point-to-point This is the initialization process for a switched point-to-point connection: 1. System z FICON technical description 15 .channel to control unit initialization 1 Also called the Hardware System Area. Channel initialization PLOGI/FLOGI QSA RNID PLOGI / LIRR / RNID ELP / LPE CU initialization PLOGI/FLOGI RNID FICON Director PLOGI / LIRR / RNID Control Unit System z10 PRLI (zHPF support) ELP / LPE Figure 2-1 FICON initialization . Process Login (PRLI) is used when channel and control unit support System z High Performance FICON (zHPF). Channel State-Change Registration (SCR) is sent to the control unit. 2. Chapter 2.Point-to-point This is the initialization process for a point-to-point connection: 1. RNID function provides specific neighbor node information. LIRR channel to the management server. 4. The channel will also login to the defined control unit N_Port. Establish Logical Path (ELP) to the control unit images that are stored in the Hardware Storage Area (HSA1) when each channel image is initialized. 3. 5. PRLI used when channel and control unit support zHPF. 5. Channel Link-Incident-Record Registration (LIRR). Channel N_Port Login process. Query Security Attributes (QSA) determines if a cascaded switch configuration is supported and if two-byte destination addresses can be used. Channel F_Port Login process. LIRR to control unit. it is a protected area of storage where control blocks used by the channel subsystem are located. Request Node Identifier (RNID) function to provide specific neighbor node information. The only difference is that the FICON channel and control unit log into (F_Port Login) two different FICON Directors. 6. 4. The channel performs this for each control unit N_Port link address defined on this channel path. Channel SCR is sent to the fabric controller. Channel N_Port Login. – – – – RNID to control unit. 2. 3. Cascaded FICON Directors The cascaded FICON Directors login process is similar to that of the switched point-to point process.

the N_Port or F_Port is in the active state. FICON Express8 will auto-negotiate to 8 Gbps.4 (FC-GS-4) standard (INCITS (ANSI) T11 group). A transport mode I/O operation uses a Transport Control Word (TCW) to send commands and data over the FC link and does not use CCWs. The channel and control unit must support zHPF. the channel program must be created and passed to the Channel Subsystem (CSS). When the SSCH is passed to the CSS (see Figure 2-2 on page 17) and the FICON channel is selected. FICON Express4 will auto-negotiate to 4 Gbps. and it is the same for all channels belonging to the platform.1. The link speed is also determined during this process. After link initialization is complete for an N_Port or F_Port. 2 Gbps.Link initialization is described in FC-FS-2. the port is considered to be operational as long as it remains in the active state. The Input/Output Supervisor (IOS) is a z/OS component. FICON registration System z10 servers support platform and name server registration to the fabric. 2.Generic Services . – Classes of service supported by the channel. – FC-4 types supported. When registered. and vendor-specific data from the node descriptor. between the CSS and the application program. The application program uses an access method that calls the I/O driver to build the channel program. FICON features support the following speeds: FICON Express2 will auto-negotiate to 2 Gbps or 1 Gbps. The FICON link will start at the highest speed and work lower until both sides choose the same speed. – Port type (N_Port_ID). 16 FICON Planning and Implementation Guide . 4 Gbps.The name includes Vendor ID. information about the channels connected to a fabric will be available to other nodes or SAN managers. or 1 Gbps. that issues the start subchannel command.2 FICON I/O request Before the actual I/O operation starts and FC frames are transmitted over the fiber cable. Channel information – World Wide Port Name (WWPN). through auto-negotiation. or channel commands when operating in command mode. – Platform type (host computer). or 2 Gbps. The attributes that are registered for the System z10 servers include: Platform information – World Wide Node Name (WWNN) . Transport mode is used by the System z High Performance FICON (zHPF) architecture. When link initialization is complete. – Platform name . The channel program may use transport commands when operating in transport mode. Platform registration is a service defined in the Fibre Channel . Command mode is used by the existing FICON architecture where an I/O operation consists of CCWs and data. product ID. the FC architecture and FICON protocols are used to build the frames required to send over the FC link.This is the node name of the platform.

when the channel receives an initiative to start an I/O operation. An application or system component invokes an I/O request. 6. IOS issues a Start Subchannel (SSCH) instruction with the Subsystem Identification word (SSID) representing the device and ORB as operands. When the 2 The UCB is a control block in memory that describes an I/O device to the operating system. Device-level information is transferred between a channel and a control unit in SB-3 or SB-4 Information Units (IUs). 2. 7. The Input Output Supervisor (IOS) will service the request from the UCB on a priority basis. 4. The request is queued on the Unit Control Block (UCB)2.Application z/OS I/O Requests SSCH (ORB) I/O Interrupt IOS UCB System z Channel Subsystem FICON Channel FC4 (Protocol) FC3 (Services) FC2 (Framing) FC1 Encode/ Decode FC0 (Optics) FC-FS FC-2 frames. protocol type = FC-SB-3 or SB-4 (FICON) SOF FC-2 Header FICON payload FC-2 CRC EOF FC Port addressing Fibre Channel Fabric FICON CU FC-2 Hdr . The application or access method provides CCWs and additional parameters in the Operation Request Block (ORB). The channel fetches from storage the Channel Command Words and associated data (for Write Operations). For example. 5. The ORB contains start-specific control information. Chapter 2. and also indicates the starting address of the Channel Programs (CPA) Channel Command.S_ID D_ID S_ID D_ID D_ID S_ID Figure 2-2 FICON channel I/O request A FICON I/O request flow is as follows: 1. 3. Information units are transferred using both link-level and device-level functions and protocols. System z FICON technical description 17 . the device-level functions and protocols obtain the command and other parameters from the current CCW and insert them into the appropriate fields within a command IU. It indicates whether the channel is operating in transport mode (zHPF support) or command mode. The CSS selects the most appropriate channel and passes the I/O request to it. The channel assembles the required parameters and fields of the FC-2 and FC-SB-3 or FC-SB-4 for the I/O request and passes them to the Fibre Channel adapter (which is part of the FICON channel).

and the payload part (the FC-SB-4 or FC-SB-3 parts). FICON frame layout The FC-4 defines the mapping rules used by FICON. FC-SB-3 and FC-SB-4 are based on the FC-4 Information Unit constructs. The FC-2 header part of the frame holds the source FC port address (S_ID) and the destination FC port address (D_ID). IUs contain device-level commands. The Fibre Channel (FC-2) sends the SB-3 and SB-4 IUs in FC-2 frames. or control information or link-level control information. link-level functions and protocols provide additional information (for example. the Control Unit Image address. data. and the Device Unit Address (UA). address identifiers and exchange ID in the frame header) and then coordinate the actual transmission of the frame on the channel path. Information associated with the execution of an I/O operation and the operation of a device is transferred between the channel and control unit as Information Units. The Fibre Channel adapter builds the complete FC-FS FC-2 serial frame and transmits it into the Fibre Channel link. As part of building the FC-FS FC-2 frame for the I/O request.command IU is ready for transmission. the FICON channel in FICON native (FC) mode constructs the 24-bit FC port address of the destination N_Port of the control unit. and the control unit image and device address within the physical CU. The receiving end node (FC adapter at the CU or another channel) rebuilds the incoming FC frames for a given IU into an inbound IU. The FC-SB-3 or FC-SB-4 header part of the frame holds the channel image address. When FC Header indicates SB-3 or SB-4 Payload: Contains channel command and/or data FICON Fibre Channel standard frame format Word 0 1 2 3 4 5 Bits 31-24 R_CTL 4 Bytes Start of Frame (SOF) 2112 Bytes Payload 24 Bytes FC Header 64 Bytes Optional Header 4 Bytes CRC Error Check 4 Bytes End of Frame (EOF) 2048 Byte Payload Bits 23-16 bits 15-8 Destination ID (Address) Source ID (Address) Sequence Count bits 7-0 Sequence ID Originator Exchange ID Responder Exchange ID R_CTL indicates if it is a Link control frame Data frame (FICON ) Exchange ID Command mode (exchange pair is is used in an active I/O operation) Figure 2-3 FICON frame layout 18 FICON Planning and Implementation Guide . status. The FC header and its content determines how the frame is to be handled within the fabric The frame contains two parts: the FC-FS FC-2 header part. The maximum amount of requester information that can fit into one FC frame is 2112 bytes. 8. As many frames will be sent to transfer an IU as it takes to hold the IU data. Figure 2-3 illustrates the FICON frame layout.

The FC header and its content determine how the frame is to be handled. The R_CTL field indicates the type of frame it is, as follows: Link Control – Acknowledge – Link Response – Link Command Data Frame – Video_Data – Link Data • Basic Link Service (BLS) • Extended Link Service (ELS) – Device_Data (Type - IP, IPI-3, SCSI, FCP, SB-3 or SB-4) • SB-3 and SB-4 is used for FICON The FC header also contains the Exchange ID. In FICON command mode, two exchange IDs are used and are called an exchange pair. Command mode uses one exchange pair for one I/O operation. An open exchange represents an I/O operation in progress over the channel. A detailed description of the FC frame and FICON FC-SB-3 or SB-4 usage documentation can be obtained from the following Web site: http://www.t11.org

2.1.3 Command mode
The System z FICON architecture defines the protocol for CCW and data pipelining, and how FICON channels operate (known as command mode). Command mode uses the CCWs and data as shown in Figure 2-4 to perform I/O operations.

SSCH I/O Request

FICON Channel
zArch CCW1 FC-SB-3 Cmd/Data IU Cmd/Data IU Cmd IU Cmd IU Cmd IU Cmd/Data IU Cmd/Data IU Data IU FC-FS FC2 FC2 FC2 FC2 FC2 FC2 FC2

Control Unit
Frame(s) Frame(s) Frame(s) Frame(s) Frame(s) Frame(s)
CCW1 CCW2 CCW3 CCW4 CCW5 CCW6 CCWn CMD2 CMR CMD1

Device

Data from CCW2 Memory CCW3
CCW4 CCW5 CCW6

CMD1 End CMD2 End

Data to Memory

CCWn

Frame(s) Frame
IU CE/DE (CCW-y)

CMDy

CMD-y End

I/O Status IU Request Control (StaAcc) Complete I/O Interrupt

Frame

Figure 2-4 FICON command mode operation

A fundamental difference with ESCON is the CCW chaining capability of the FICON architecture. ESCON channel program operation requires a Channel End/Device End (CE/DE) after executing each CCW. FICON supports CCW chaining without requiring a CE/DE at the completion of each CCW operation.

Chapter 2. System z FICON technical description

19

The ESCON channel transfers the CCW to the control unit and waits for a CE/DE presented by the control unit after execution of the CCW by the device (CCW interlock). After receiving CE/DE for the previous CCW, the next CCW is transferred to the control unit for execution. With a FICON channel, CCWs are transferred to the control unit without waiting for the first command response (CMR) from the control unit or for a CE/DE after each CCW execution. The device presents a logical end to the control unit after each CCW execution. After the last CCW of the CCW chain has been executed by the CU/device, the control unit presents a CE/DE to the channel. In addition, FICON channels can multiplex data transfer for several devices at the same time. This also allows workloads with low to moderate control unit cache hit ratios to achieve higher levels of activity rates per channel.

2.1.4 Transport mode
The way transport mode handles an I/O operation is significantly different from the CCW operation for command mode. While in command mode, each single CCW is sent to the control unit for execution. In transport mode, all commands are sent in a single frame to the control unit. The FICON channel uses transport mode if the control unit supports zHPF. When operating in transport mode, an I/O operation will use a transport control word (TCW) to send commands and data, as illustrated by Figure 2-5.

FICON Channel
zHPF Arch FC-SB-4 FC-FS

Control unit

SSCH I/O Request Data from Memory

4 KB read cmds
Prefix + 64 byes of data + Read cmd (4k) + Read cmd (4k) + Read cmd (4k) + Read cmd (4k) Cmd/Data IU FC2

Frame(s)

Data to Memory I/O Request Complete

TCW is used to obtain cmds and data for I/O operation

Frame(s)

Data (16 KB) + CRC

Frame(s)

Status
Total frames 11 20

I/O operation with four 4 KB reads Transport mode Command mode

Total commands 5 5

Total exchanges 1 2 (exchange pair)

Figure 2-5 FICON transport mode operation

The FICON Express8, FICON Express4, and FICON Express2 features support transport mode. A parameter in the Operation Request Block (ORB) is used to determine how the FICON channel will operate (command mode or transport mode). The mode used for an I/O operation also depends on the settings in the z/OS operating system.

20

FICON Planning and Implementation Guide

The application or access method provides the channel program commands and parameters in the ORB. Bit 13 in word 1 of the ORB specifies how to handle the channel program in either command mode or transport mode.

zHPF support determination
zHPF support is determined at link initialization. Figure 2-6 shows the initialization steps used to determine if transport mode will be used. During link initialization, both the channel and the control unit indicate whether they support zHPF. In the response to the Request Node ID (RNID) Extended Link Sequence (ELS), the Process Login (PRLI) support indicator is presented. In the RNID response, Process Login (PRLI) support is indicated in the node parameters, bit three. When bit three is one (1): – The node supports the FC-SB-4 process login using the PRLI extended link service. • FC-SB-4 protocol is used by zHPF When bit three is zero (0): – The node does not support the FC-SB-4 process login. If PRLI is supported, the channel sends a PRLI ELS. The PRLI response then indicates that zHPF is supported by the control unit.

Channel-to-CU initialization PLOGI/ LIRR Send RNID Accept RNID (Indicate support for FC-SB-4 process login) Perform PRLI and determine transport mode support)

Control Unit

Figure 2-6 FICON sequence to determine zHPF support

A channel that supports the process login (PRLI) extended link service and transport mode operations will send a request to each control unit in its configuration that also supports the process login ELS. The PRLI is used to determine whether the control unit supports transport mode operations. zHPF provides additional capabilities such as being able to interrogate the CU before an actual missing interrupt occurs. Because zHPF uses TCWs (which do not time the channel operation in the same way as command mode), a mechanism must be used to prevent unnecessarily invoking the Missing Interrupt Handler (MIH) recovery actions. Transport mode provides an in-band method for z/OS to query the state of the I/O at the control unit without invoking error recovery or retry.

Missing Interrupt Handler
In channel operations, there is a Missing Interrupt Handler (MIH) time value set for various device classes. The MIH is constantly polling and is used to determine if there are any I/O operations that are not completing. For each of these I/O requests that are not completing, a recovery action will be performed. The recovery action could be one or more of the following: Issue a message Generate a logrec record for diagnostic purposes Terminate the I/O request Requeue the I/O request
Chapter 2. System z FICON technical description

21

Normally when a channel does not get a CMR response from a started I/O operation within the MIH value, it invokes MIH recovery procedures, which can then terminate the I/O in progress. The transport mode operation called interrogate is used to query the state of the CU. (Note that the interrogate process does not affect the state of the primary operation, and is sent before the MIH value is exceeded.) The CU provides an interrogate response IU that contains extended status that describes the state of the primary operation. z/OS can then decide whether recovery is needed, or to reset MIH value for this I/O operation.

2.1.5 Modified Indirect Data Address Word
On System z, the Modified Indirect Data Address Word (MIDAW) provides alternatives to using CCW data chaining in channel programs. The MIDAW facility has been added to z/Architecture and can coexist with the current CCW IDAW facility. MIDAW allows scattering of data in memory for non-contiguous real pages. (This is sometimes known as scatter-read or scatter-write.) Although the CCW IDAW function requires all but the first and last IDAW in a list to deal with complete 2 KB or 4 KB units of data, the MIDAW facility allows page boundary crossing on either 2 KB or 4 KB boundaries. This allows access to data buffers anywhere in a 64-bit buffer space. Figure 2-7 illustrates an example of MIDAW usage.

CCW with IDAW flag set (with ORB specifying 4K blocks) CCWs
cmd 04

IDAW address

Real address IDAWs Real address Real address

IDAWs usage limited because of addressing requirements

CCW with MIDAW flag set CCW
cmd 01

MIDAW address

MIDAWs

reserved reserved reserved L

2k 32 1k

Real address Real address Real address

IDAWs cannot be used for scattered reads or writes

Figure 2-7 Command mode MIDAWs

2.1.6 Transport Indirect Data Address Word (TIDAW)
When the System z channel is operating in transport mode, Transport Indirect Data Address Work (IDAW) is used. TIDAWs and MIDAWs are similar in concept and provide the capability of scattered reads and writes.

22

FICON Planning and Implementation Guide

TIDAWs are used when certain flag bits are set in the transport control word. Figure 2-8 illustrates an example of TIDAW usage.

TCW flag set for TIDAWS
TIDAWS

flags reserved count real address flags reserved count real address flags reserved count real address

Figure 2-8 Transport mode TIDAWs

TIDAWs and MIDAWs are used with z/OS extended format data sets that use internal structures (usually not visible to the application program) that require scatter-read or scatter-write operation.

2.1.7 Open exchange
An open exchange is part of FICON (and FC) terminology. Many I/O operations can be in progress over FICON channels at any one time. For example, a disk I/O operation might temporarily disconnect from the channel while performing a seek operation or while waiting for a disk rotation. During this disconnect time, other I/O operations can be managed, as explained here: Command mode open exchanges In command mode, the number of open exchanges is limited by the FICON Express feature. FICON Express8, FICON Express4, FICON Express2 allow up to 64 open exchanges. One open exchange (actually, this is an exchange pair) in command mode is the same as one I/O operation in progress. Transport mode open exchanges In transport mode, one exchange is sent from the channel to the control unit. Then the same exchange ID is sent back from the control unit to the channel to complete the I/O operation. The maximum number of simultaneous exchanges the channel can have open with the CU is up to 750 exchanges. The CU sets the maximum number of exchanges in the status area of the transport mode response IU (the default number is 64 and can be increased or decreased).

2.1.8 Buffer-to-buffer credit usage in FICON
Normally buffer credits are only an issue for extended distances. The buffer-to-buffer value determines the distance that the two nodes can be apart and still maintain the supported link data rate. All System z FICON channel features have enough buffer credits to support a distance of 10 km. When FICON frames travel over the link, a flow control mechanism must be in place to ensure the transmitter does not overrun the receiver. FICON mainly delivers class 3 type frames that use buffer-to-buffer credits to manage the flow control Ports have buffers for the temporary storage of frames. Each buffer can contain up to 2112 bytes of data. Buffers are referred to as credits. A credit represents a receiving port's ability to

Chapter 2. System z FICON technical description

23

accept one frame. At port initialization (login), buffer credit values are exchanged between two ports based on the number of buffers available for the ports. For example, the FICON Express8 feature has 40 buffer credits. This means that the FICON Express8 receiver is capable of storing up to 40 frames of data at any one time. The other FICON features support the following: FICON Express4 contains 212 buffer credits. FICON Express2 contains 107 buffer credits. FICON Express contains 64 buffer credits. For a more complete explanation about how buffer credits work, refer to “Buffer credits” on page 39.

2.1.9 Extended distance FICON
Degradation of performance at extended distances can be avoided by implementing an enhancement to the industry standard FICON architecture. The enhancement is a protocol for persistent IU pacing, and it can be used when the FC-SB-4 process login is supported. Control units that exploit the architecture can increase the pace count (the number of IUs allowed to be underway between the channel and the control unit). The IU pacing protocol controls the number of IUs that can be in flight from a channel to a control unit. The control unit may increase the pacing count (the number of IUs allowed to be in flight from a channel to a control unit) in the first Command Response IU sent to the channel. When the enhancement for persistent IU pacing is not used, the increased pacing count is valid only for the remainder of the current outbound exchange; see Figure 2-9.

FICON channel FC–4 IU (command / data) FC – 2 frames
frame frame frame frame frame

CU
CU response - I got frames send more IUs (CU may modify IU count at this time)

Figure 2-9 IU pacing for extended distance

At the start of every I/O, the IU count is reset to the default number. The IU pacing protocol, as defined, has the limitation that the first burst of IUs from the channel to the control unit may be no larger than a default value of 16. This causes a delay in the execution of channel programs with more than 16 IUs at long distances because a round trip to the control unit is required before the remainder of the IUs can be sent by the channel, upon the receipt of the first command response, as allowed by the increased pacing count. A channel may operate in the default IU pacing mode or in the persistent IU pacing mode. During initialization, the specific node descriptor information is exchanged between the channel and control unit. This information includes SB-4 support and will indicate whether the node supports concurrent enablement of the persistent IU pacing function. When a channel that supports concurrent enablement of the persistent IU pacing function receives a node descriptor from a control unit with bit 7 of byte 1 equal to one (1), the channel enables persistent IU pacing for all currently established logical paths with the control unit. When a control unit that supports concurrent enablement of the persistent IU pacing function receives a node descriptor from a channel with bit 7 of byte 1 equal to one (1), the 24
FICON Planning and Implementation Guide

Channel Initialization – Establish logical path (ELP) LP established – and increase IU count (Persistent IU pacing) CU Send FF IUs FICON channel FC–4 IU (command / data) FC – 2 frames frame frame frame frame frame frame frame frame frame CU response . 2. this feature is only supported on System z10 and System z9 severs if the feature was carried forward on an upgrade. This may improve performance of long I/O programs at higher link speeds and long distances by allowing the channel to send more IUs to the control unit and eliminating the delay of waiting for the first Command Response. Chapter 2. FCP.I got frames (CU may modify IU count at this time) CU DS8000 Figure 2-10 Persistent IU pacing for extended distance Support is provided on the IBM System Storage™ DS8000® with the appropriate licensed machine code level. The channels residing on a FICON Express feature may be defined for FC. or FICON Express2 feature can be configured individually. The channels residing on a single FICON Express8.control unit enables persistent IU pacing for all currently established logical paths with the channel. For logical paths that are established subsequent to the processing of the node descriptor. and is exclusive to System z10 servers. however.2 System z FICON feature support Table 2-1 on page 26 lists the FICON features available on the System z10 and System z9 servers. Some features are only available when carried forward on a server upgrade and FICON Express8 features are only available on the System z10 server. and can be defined in different channel modes (FC or FCP). FICON Express4. Persistent IU pacing is a method for allowing a channel and control unit supporting the FC-SB-4 process login to retain a pacing count that can be used at the start of execution of a channel program. System z FICON technical description 25 . and the number of FICON channels support for each. and uses that pacing count value as its new default pacing count for any new channel programs issued on the same logical path. Figure 2-10 illustrates an example of using persistent IU pacing for extended distance. or FCV channel mode. the persistent IU pacing bit in the ELP/LPE IU optional features field is used to enable or disable persistent IU pacing. presented by the control unit in accordance with the standard. The channel retains the pacing count value.

and printers. withdrawn from marketing for System z10 servers FICON Express8 The FICON Express8 features are exclusive to System z10 and are designed to deliver increased performance compared to the FICON Express4 features. Carry forward on an upgrade b.Table 2-1 System z server FICON feature support Maximum number of channels Channel feature Feature codes 2319 2320 3319 3320 3321 3322 3318 3323 3324 3325 3326 z9 EC 120 120 336 336 336 336 n/a n/a 336 n/a n/a z9 BC R07/S07 32/40 32/40 64/80a 64/80a 64/112 64/112 32/56 32/56 64/112 n/a n/a z10 EC 120 120 336a 336a 336 336 n/a n/a 336 336 336 z10 BC 40 40 112a 112a 128 128 64 64 128 128 128 Channels per feature 2 2 4 4 4 4 2 2 4 4 4 Channel increments (orderable) 2 2 4 4 4 4 2 2 4 4 4 FICON Express LXa FICON Express SXa FICON Express2 LX FICON Express2 SX FICON Express4 10KM LXb FICON Express4 SX b FICON Express4-2C SX FICON Express4-2C 4KM LX FICON Express4 4KM LXb FICON Express8 10KM LX FICON Express8 SX a. z/TPF. All FICON Express8 features use Small Form Factor Pluggable (SFP) optics to permit each channel to be individually serviced in the event of a fiber optic module failure. z/VSE™. The FICON Express8 features are designed for connectivity to servers. disks. and each feature occupies a single I/O slot. and Linux® on System z environments CHPID type FCP – Fibre Channel Protocol traffic for communication with SCSI devices – Supported in z/VM. 4 Gbps. utilizing one CHPID per channel. The FICON Express8 features have four independent channels. High Performance FICON for System z (zHPF). This concurrent update capability allows you to continue to run workloads through other channels while the FICON Express8 features are being added. The FICON Express8 features are ordered in 4-channel increments and are designed to be added concurrently. switches. tapes. z/VM®. z/VSE. and they can be defined as: CHPID type FC – Native FICON. Directors. A link data rate of 1 Gbps is not supported. and Linux on System z environments 26 FICON Planning and Implementation Guide . 2009. FICON Express8 CHPIDs can be defined as a spanned channel and can be shared among logical partitions within and across CSSs. Each channel supports 2 Gbps. The traffic on the other channels on the same feature can continue to flow if a channel requires servicing. and 8 Gbps link data rates with auto-negotiation. and FICON Channel-to-Channel (CTC) traffic – Supported in the z/OS. Effective October 27.

High Performance FICON for System z (zHPF).5/125 micrometer multimode fiber optic cable terminated with an LC Duplex connector. and FICON Channel-to-Channel (CTC) traffic – Supported in the z/OS. Directors. and each feature occupies a single I/O slot.5 or 50 micron) fiber cable can be used with the FICON Express4 10KM LX feature. All FICON Express4 features use Small Form Factor Pluggable (SFP) optics to permit each channel to be individually serviced in the event of a fiber optic module failure. and printers.2 miles) at 8 Gbps. Each channel supports 9 micron single mode fiber optic cable terminated with an LC Duplex connector. System z FICON technical description 27 . FICON Express4 The FICON Express4 features have four (or two. and Linux on System z environments CHPID type FCP – Fibre Channel Protocol traffic for communication with SCSI devices – Supported in z/VM. for the 2-port features) independent channels. This concurrent update capability allows you to continue to run workloads through other channels while the FICON Express4 features are being added. The traffic on the other channels on the same feature can continue to flow if a channel requires servicing.FICON Express8 10KM LX Feature code 3325 is designed to support unrepeated distances up to 10 km (6.5 miles) or the link loss budget between devices exceeds 2 dB. Each channel supports 1 Gbps. Multimode (62. z/VSE. utilizing one CHPID per channel. FICON Express4 CHPIDs can be defined as a spanned channel and can be shared among logical partitions within and across CSSs. and 4 Gbps link data rates with auto-negotiation. z/VSE. The use of multimode cable types requires a mode conditioning patch (MCP) cable. Each channel supports 50/125 micrometer multimode fiber optic cable or a 62. This feature should be used when the unrepeated distance between devices is greater than 4 km (2. The FICON Express4 features are designed for connectivity to servers. The FICON Express4 features are ordered in 4-channel (or 2-channel) increments and are designed to be added concurrently. tapes. and Linux on System z environments FICON Express4 10KM LX Feature code 3321is designed to support unrepeated distances up to 10 km (6. switches. disks. z/VM. z/TPF. A 10KM LX transceiver is designed to interoperate with a 10KM LX transceiver. Interoperability of 10 km Chapter 2. and they can be defined as: CHPID type FC – Native FICON.2 miles) at 4 Gbps. 2 Gbps. FICON Express8 SX Feature code 3326 is designed to support unrepeated distances up to 150 meters (492 feet) at 8 Gbps. Each channel supports 9/125 micrometer single mode fiber optic cable terminated with an LC Duplex connector.

is designed to support unrepeated distances up to 270 meters (886 feet) at 4 Gbps. Each channel supports 9 micron single mode fiber optic cable terminated with an LC Duplex connector.transceivers with 4 km transceivers is supported if the unrepeated distance does not exceed 4 km. The use of multimode cable types requires a mode conditioning patch (MCP) cable. FICON Express4 4KM LX Feature code 3324 is designed to support unrepeated distances up to 4 km (2. The use of multimode cable types requires a mode conditioning patch (MCP) cable. Note: FICON Express4-2C SX is only available on System z10 BC and System z9 BC servers. FICON Express4 SX Feature code 3322 is designed to support unrepeated distances up to 270 meters (886 feet) at 4 Gbps. FICON Express4-2C 4KM LX Feature code 3323. FICON Express2 The FICON Express2 SX and FICON Express2 LX features have four independent channels.2 miles) transceivers is supported if the unrepeated distance does not exceed 4 km. Interoperability of 4 km transceivers with 10 km (6. Multimode(62.5 micron or 50 micron multimode fiber optic cable terminated with an LC Duplex connector.2 miles) transceivers is supported if the unrepeated distance does not exceed 4 km.5 or 50 micron) fiber cable can be used with the FICON Express4-2C 4 KM LX feature. Each channel supports 62. with two channels per feature. with two channels per feature. Each channel supports 62. Note: FICON Express4-2C 4KM LX is only available on System z10 BC and System z9 BC servers. Interoperability of 4 km transceivers with 10 km(6. utilizing one CHPID per channel and four 28 FICON Planning and Implementation Guide .5 or 50 micron) fiber cable can be used with the FICON Express4 4 KM LX feature.5 miles) at 4Gbps.5 miles) at 4 Gbps. Each channel supports 9 micron single mode fiber optic cable terminated with an LC Duplex connector. with each feature occupying a single I/O slot. FICON Express4-2C SX Feature code 3318.5 micron or 50 micron multimode fiber optic cable terminated with an LC Duplex connector. is designed to support unrepeated distances up to 4 km (2. Multimode (62.

and FICON Channel-to-Channel (CTC) traffic – Supported in the z/OS. High Performance FICON for System z (zHPF). FICON Express2 LX Feature code 3319 is designed to support unrepeated distances up to 10 km (6. and Linux on System z environments CHPID type FCV Chapter 2. and printers. Each channel supports 62. The FICON Express2 SX and LX features are ordered in 4-channel increments and designed to be added concurrently. Each channel can be configured individually and support 1 Gbps link speed. Each channel supports 9 micron single mode fiber optic cable terminated with an LC Duplex connector. z/VM. tapes. z/TPF. FICON Express2 CHPIDs can be defined as a spanned channel and can be shared among logical partitions within and across CSSs. FICON Express2 SX Feature code 3320 is designed to support unrepeated distances up to 500 meters (1640 feet) at 2 Gbps.CHPIDs per feature. z/VSE. Multimode (62. and Linux on System z environments CHPID type FCP – Fibre Channel Protocol traffic for communication with SCSI devices – Supported in z/VM. and they can be defined as: CHPID type FC – Native FICON. switches. This concurrent update capability allows you to continue to run workloads through other channels while the FICON Express2 features are being added. System z FICON technical description 29 . z/TPF. The use of multimode cable types requires a mode conditioning patch (MCP) cable.5 or 50 micron) fiber cable can be used with the FICON Express2 LX feature. Directors. FICON Express The two channels residing on a single FICON Express feature occupy one I/O slot in the System z I/O cage. z/VSE. switches. disks. The FICON Express features are designed for connectivity to servers. The FICON Express2 features are designed for connectivity to servers. tapes. The link speed is auto-negotiated.5 micron or 50 micron multimode fiber optic cable terminated with an LC Duplex connector. and Linux on System z environments Note: FICON Express2 is supported on System z10 and System z9 servers only if carried forward on an upgrade. while continuing to support 1 Gbps and 2 Gbps link data rates. and Linux on System z environments CHPID type FCP – Fibre Channel Protocol traffic for communication with SCSI devices – Supported in z/VM. and printers.2 miles) at 2 Gbps. and they can be defined as: CHPID type FC – Native FICON and FICON Channel-to-Channel (CTC) traffic – Supported in the z/OS. Directors. z/VSE. z/VM. disks. z/VSE.

5 µm 2320 FICON Express SX LC Duplex MM 50 µm 1 Gbps 1 or 2 Gbpsa 30 FICON Planning and Implementation Guide . Each channel supports 62. z/VSE. and Linux on System z environments Note: FICON Express is supported on System z10 and System z9 servers only if carried forward on an upgrade. Table 2-2 FICON channel specifications Feature code Feature name Connector type Cable type Link data rate 1 or 2 Gbpsa Server z10 ECb z10 BCb z9 ECb z9 BCb z10 ECb z10 BCb z9 ECb z9 BCb z10 ECb z10 BCb z9 ECb z9 BCb SM 9 µm 2319 FICON Express LX LC Duplex with MCPc: MM 50 µm or MM 62.2 miles) at 1 Gbps. Each channel supports 9 micron single mode fiber optic cable terminated with an LC Duplex connector.– Supported in the z/OS.5 or 50 micron) fiber cable can be used with the FICON Express LX feature. FICON Express8 features do not support the attachment of MCP cables. z/TPF. 5µm MM 62.5 micron or 50 micron multimode fiber optic cable terminated with an LC Duplex connector. Multimode (62. The use of multimode cable types requires a mode conditioning patch (MCP) cable. Table 2-2 summarizes the available FICON feature codes and their respective specifications. FICON Express SX Feature code 2320 is designed to support unrepeated distances up to 860 meters (2822 feet) at 1 Gbps. FICON Express LX Feature code 2319 is designed to support unrepeated distances up to 10 km (6. Notes: Mode Conditioning Patch (MCP) cables can be used with FICON features that can operate at a link data rate of 1 Gbps (100 MBps) only. z/VM.

or 4 Gbpsa z10 ECb z10 BCb z9 EC z9 BC z10 EC z10 BC z10 EC z10 BC 3325 FICON Express8 LC Duplex 10KM LX 2. or 4 Gbpsa z10 ECb z10 BCb z9 EC z9 BC z10 ECb z10 BCb z9 EC z9 BC 1. 4. or 4 Gbpsa 1.5 µm 3318 FICON Express4-2C SX LC Duplex MM 50 µm SM 9 µm 3321 FICON Express4 10KM LX LC Duplex with MCPc : MM 50 µm or MM 62. Only supported when carried forward on an upgrade c. or 4 Gbpsa 1. Supports auto-negotiate with neighbor node b. Mode conditioning patch cables may be used d. 2.5 µm MM 62. 2.5 µm SM 9 µm MM 62. Chapter 2. 2. Supports auto-negotiate with neighbor node for link data rates 8. 2. or 8 Gbpsd 2. 2.5 µm 3322 FICON Express4 SX LC Duplex MM 50 µm SM 9 µm 3323 FICON Express4-2C 4KM LX LC Duplex with MCPc : MM 50 µm or MM 62. System z FICON technical description 31 .Feature code Feature name Connector type Cable type Link data rate 1 or 2 Gbpsa Server z10 ECb z10 BCb z9 ECb z9 BCb z10 ECb z10 BCb z9 ECb z9 BCb z10 ECb z10 BCb z9 ECb z9 BCb z10 BC z9 BC SM 9 µm 3319 FICON Express2 LX LC Duplex with MCPc : MM 50 µm or MM 62.5 µm SM 9 µm 3324 FICON Express4 4KM LX LC Duplex with MCPc : MM 50 µm or MM 62. 4.5 µm 3326 FICON Express8 SX LC Duplex MM 50 µm 1. or 8 Gbpsd a.5 µm fiber optic cabling in the same physical link. and 2 Gbps only Note: IBM does not support a mix of 50 µm and 62. 4. or 4 Gbpsa z10 BC z9 BC 1.5 µm 3320 FICON Express2 SX LC Duplex MM 50 µm 1 Gbps 1 or 2 Gbpsa MM 62.5 µm MM 62.

Refer to Table 4-1 on page 67 for the allowable maximum distances and link loss budgets based on the supported fiber optic cable types and link data rates. 32 FICON Planning and Implementation Guide .

a Director typically means more fully redundant hardware resulting in higher availability. However. compared to a switch. All rights reserved. It also offers information regarding FICON Directors that have been qualified by IBM for System z use. Terminology: Often the terms Director and switch are used interchangeably.3 Chapter 3. 2009. FICON Director technical description This chapter provides a technical description of the common components and functions of a FICON Director that are needed to support FICON environments. larger port capacity. and additional capabilities. 2005. 2006. © Copyright IBM Corp. 33 .

1 Switched point-to-point configuration In a switched point-to-point connection. The FICON Director provides the ability to logically connect multiple control units to a set of channels from one or more System z servers. or to different FICON control units. Between the FICON Director (F_Port) and the control unit (N_Port) 34 FICON Planning and Implementation Guide . in conjunction with the FICON Director. control units.3. or System z server (channel-to-channel). printer. Figure 3-1 shows a conceptual view of frame processing in a FICON Director with multiple FICON channels and control units. Cascaded FICON Directors (through two FICON Directors to FICON-capable control units) 3. See Chapter 2.frame processing The FICON channel supports multiple concurrent I/O connections. “System z FICON technical description” on page 13 for more details. FICON Channel FIC ON Fra m es O FIC es r am NF FICON Control Unit es FIC ON F ram es FICON Director ram NF ICO F es ram NF O FIC s me Fra ON FIC FIC O FIC ON Fra me s NF ram es FICON Channel FICON Control Unit Figure 3-1 FICON Director . such as a storage device. Its primary function is to allow the dynamic connections between System z servers. A FICON channel. One might say that the FICON Director is at the heart of I/O connectivity for System z servers. Between the System z (N_Port) and the FICON Director (F_Port) 2. at least two Fibre Channel (FC) links are needed in the channel-control unit path: 1. A FICON channel uses the Fibre Channel communication infrastructure to transfer channel programs and data through its FICON features to another FICON-capable node.1 The role of the FICON Director The FICON Director provides a variety of features and functions.1. which in turn minimizes the cabling infrastructure and maximizes the connectivity. can operate in two topologies: 1. Switched point-to-point (through a single FICON Director to FICON-capable control units) 2. Each concurrent I/O operation can be to the same FICON control unit (but to different devices/control unit images). or another FICON Director.

1. or an interconnection of two FC links through a FICON Director. It does so by logging into the fabric using fabric login (FLOGI ELS). Chapter 3. FICON Director technical description 35 . that provide the physical transmission path between a channel and a control unit. Multiple channel images and multiple control unit images can share resources of the FC link and FICON Directors.The FICON channel determines whether the associated link is in a point-to-point or switched topology. Between the two FICON Directors (E_Ports) 3. the physical paths are the FC links. Sharing a control unit through a FICON Director means that communication from a number of System z channels to the control unit can take place over one Director. The communication path between a channel and a control unit is composed of two different parts: the physical channel path and the logical path. such that multiplexed I/O operations can be performed. An example of cascaded FICON Director topology is shown in Figure 3-3 on page 36. also known as Inter-Switch Links (ISLs). System z Server Storage Determines fabric connection (point-to-point or switched) Fabric Login (FLOGI) FICON Channel FICON Channel N_Ports F_Ports FICON Director F_Ports N_Ports FICON Control Unit Port FICON Control Unit Port Logical Path Figure 3-2 Switched point-to-point configuration Multiple channel images and multiple control unit images can share the resources of the FC link and the FICON Director. Between the System z (N_Port) and the FICON Director (F_Port) 2.2 Cascaded FICON Director configuration In a cascaded FICON Director configuration. Between the FICON Director (F_Port) and the control unit (N_Port) With this configuration. Channels and control unit links can be attached to the FICON Director in any combination. A FICON (FC-SB-3 or FC-SB-4) logical path is the relationship established between a channel image and a control unit image for communication during execution of an I/O operation and presentation of status. In a FICON switched point-to-point topology (with a single Director). at least three Fibre Channel (FC) links are needed in the channel-control unit path: 1. 3. The FLOGI-ACC (accept) response indicates if the channel (N_Port) is connected to another N_Port (point-to-point) or an F_Port (fabric port). and checking the accept response to the fabric login (ACC ELS). An example of a switched point-to-point topology is shown in Figure 3-2. A FICON channel can communicate with a number of FICON control units on different ports as well. such that multiplexed I/O operations can be performed. depending on configuration requirements and available resources in the Director. the connection between sites can consist of multiple FC links.

It typically features an N+1 redundancy scheme for all of its components. such as loading and activating firmware. depending on configuration requirements and available Director ports.Channels and control unit FC links can be attached to the FICON Directors in any combination. System z Server Determines fabric connection (point-to-point or switched) Storage Fabric Login (FLOGI) FICON Channel FICON Channel ISLs FICON Control Unit Port FICON Control Unit Port FICON Director Logical Path E_Ports FICON Director Figure 3-3 Cascaded FICON Director configuration A FICON (FC-SB-3 or FC-SB-4) logical path is the relationship established between a FICON channel image and a control unit image for communication during execution of an I/O operation and presentation of status. In a cascaded FICON Director topology. A FICON Director consists of a chassis that usually holds the following types of components: Backplane module Power supply assembly Fan module assembly I/O module Switch module 36 FICON Planning and Implementation Guide . Sharing a control unit through a FICON Director means that communication from a number of channels to the control unit can take place either over one Director-to-CU link (in the case where a control unit has only one FC link to the FICON Director). the FICON Director remains operational.1. interconnected by the Directors. or over multiple link interfaces (in the case where a control unit has more than one FC link to the FICON Director). This redundancy results in high availability. and replacement of failed parts. A FICON channel can also communicate with a number of FICON control units on different ports in the second FICON Director. 3. that provide the physical transmission path between a FICON channel and a control unit. Even in the case of failure of any single component. The communication path between a channel and a control unit is composed of two different parts: the physical channel path and the logical path. Figure 3-3 illustrates the configuration. the physical paths are the FC links. Most activities can be done non-disruptively.3 Basic components of a FICON Director A FICON Director is an expandable and scalable Fibre Channel platform designed to support FICON attachment of System z servers and storage products.

1. firmware upgrades can be performed concurrent to FICON Director operation. Due to the usage of Small Form Factor pluggable (SFP) ports instead of fixed optical port cards. If the active switch module fails or is uninstalled. Oversubscription may only occur when simultaneous activities are present and backplane capacity is depleted. Failover occurs as soon as the active switch module is detected to be faulty or uninstalled. The base configuration always contains two switch modules for redundancy reasons. port upgrades are simple and nondisruptive to other ports on the I/O module. The fan module assembly is hot-swappable. FICON Director technical description 37 . as well. It is specified in the FICON Director as a decimal number that is converted to a hexadecimal value for the Chapter 3.4 Basic functions of a FICON Director In this section we discuss the basic functions of a FICON Director that are needed to support a FICON environment. Fan module assembly Fan module assemblies provide the cooling capability for the FICON Director. 48 ports operating at 8 Gbps). Some common FICON Director terms are described. load sharing). Domain ID or switch address The Domain ID is used to define the FICON Director address (switch address). switch. Oversubscription The term oversubscription implies that the port may not be able to drive at full speed due to limited backplane capacity. The backplane gives you the capability to increase the number of ports by adding I/O modules. The active switch module performs following control functions: Switch initialization High availability and switch drivers Name server SNMP Zoning Typically. I/O module The main function of the I/O module is to provide the physical connectivity between FICON channels and control units that are being connected to the Director. providing full redundancy (and in some cases. the standby switch module automatically becomes the new active switch module.Backplane module The backplane provides the connectivity within the chassis and between all system (I/O. 3. This usually applies to I/O modules with high port counts (for example. including the power supply assemblies and fan assemblies. and control) modules. any mix of short wavelength and long wavelength ports is allowed with any I/O module. Switch module The switch module on the FICON Director contains the microprocessor and associated logic that provides overall switching in the Director. the remaining fans may increase their speeds to a higher operating rpm to compensate for the single fan failure. Each power supply assembly is a removable. hot-swappable unit. In the event of a single fan failure. Power supply assembly Usually two power assemblies are installed in a chassis. Typically.

even if it is rebooted. in many other cases they are different. port numbers and port addresses are the same. The switch address is used in the HCD or IOCP definitions. we recommend that you use the same value as the switch address and the Domain ID when referring to a FICON Director.use in the System z server. Insistent means that the FICON Director uses the same Domain ID every time. The Domain ID is assigned by the manufacturer. The two-byte link address consists of the Domain ID and the port address to which the device is connected. Port number The port number is typically used in open systems environments to refer to a port on a Director. In some cases. Port address or link address The port address is a hexadecimal number between x’00’ and x’FF’ that is used in the IOCP or HCD to define a port on the FICON Director to which the device is connected. The switch address is a hexadecimal number used in defining the Director in the System z environment. However. and it can be customized to a different value. This is known as a one-byte link address. Port addressing is platform. Some FICON Directors can modify the port address to any hexadecimal number that is required. Paths are computed by adding the cost of all the links traversed by the path. Each FICON Director in a fabric must have a unique switch address. Although the switch ID can be different than the switch address or Domain ID. and it can be any hex value between x’00’ and x’FF’. Fabric Shortest Path First Fabric Shortest Path First (FSPF) is a Fibre Channel routing protocol that applies a weighting factor to paths based on link speed and status. FSPF associates a cost for each active link and computes paths from a Director to all the other Directors in the fabric. 38 FICON Planning and Implementation Guide . depending on the Director port characteristics. Insistent Domain IDs are required in a cascaded FICON Director environment. Then a path is chosen with the lowest cost. ensure that you use values within the FICON Director’s range. The Domain ID and the switch address must be the same when referring to a Director in a System z environment. It must be unique within the fabric and insistent. There is no need for a Domain ID to ever change in a FICON environment. ISLs carry frames originating from N_Ports. The switch ID must be assigned by the user.and vendor-specific. Inter-Switch Link Inter-Switch Links (ISLs) are the fiber optic cables that connect two Directors using expansion ports (E_Ports). where it is part of the two-byte link address defined in the IOCP or HCD. When defining the switch addresses in the HCD or IOCP. Multiple ISLs are usually installed between Directors to increase bandwidth and provide higher availability. and it must be unique within the scope of the definitions (IOCP and HCD). The switch ID in the CHPID statement is basically used as an identifier or label. The valid Domain ID range for the FICON Director is vendor-dependent. and fabrics will come up faster after recovering from a failure if the Domain ID is insistent. as well as frames that are generated within the fabric for management purposes. The port address is also needed in cascaded FICON Directors.

However. Buffer credits are mainly a concern when dealing with extended distances and higher speeds. and is used by Class 2 and Class 3 traffic. Buffer credits Buffer credits are used by the Fibre Channel architecture as a flow control mechanism to represent the number of frames a port can store. It does not ensure all ISLs are equally utilized. It relies on the receiver-ready (R_RDY) signal to replenish credits. the address assigned to the N_Port performing fabric login (FLOGI) by the alternate port should be the same as the one that would have been assigned by the original port. the control unit link address is defined in the channel configuration file (IOCP) of the System z server. followed by automatic discovery of the device through the name server. the FICON Director must ensure that the N_Port address for the control unit remains the same even after the cable is reconnected to a different switch port. and includes control functions like blocking and unblocking of ports. It allows z/OS to manage a FICON Director with the same level of control and security as for an ESCON Director. and it does not inhibit congestion even if the other ISLs between Directors are not fully utilized. in a FICON environment. Host communication is via IBM Tivoli® Systems Automation for z/OS (SA for z/OS). The original port will assume the port address associated with the alternate port. a poorly designed configuration may also consume available buffer credits and have a negative impact on performance. FICON Director technical description 39 . However. Chapter 3. a failing port typically results in a cable being reconnected to another available port. After port swapping. The total number of available buffer credits (BCs) that a receiver has is determined during the link initialization process (see Figure on page 40). the CUP can provide the following: I/O management of the FICON Director FICON Director performance monitoring Error detection of the FICON Director Port swapping Port swapping refers to the capability of a FICON Director to redirect traffic on a failed F_Port to a working F_Port without requiring a change in System z I/O configuration. as well as monitoring and error reporting functions. In a Fibre Channel fabric.Note: FSPF only provides basic utilization of ISLs with the same cost. Control Unit Port The Control Unit Port (CUP) protocol is used by System z to provide in-band management for FICON Directors. Therefore. Buffer credit flow control is done at the link level (that is. Although the FICON Director is transparent to the operating system in the path to a FICON control unit or device during the execution of an I/O operation. between an N_Port and an F_Port).

4 (per km at 8 Gbps) x 10 km = 40 buffer credits 40 FICON Planning and Implementation Guide . use the following criteria: 1 buffer credit per kilometer (km) at 2 Gbps 2 buffer credits per kilometer (km) at 4 Gbps 4 buffer credits per kilometer (km) at 8 Gbps In this case. The buffer credit count value is incremented when an acknowledgment (R_RDY) is received (see Figure 3-6). BC counter Channel N_Port Xmit Rec 8 Credit 1 Frame 3 Frame 2 10 Buffers Frame 1 BC counter CU N_Port 5 Buffers 4 Xmit Frame 1 R_RDY R_RDY is returned to indicate a buffer is available for additional frame BC count decremented when frame is sent Figure 3-6 Buffer credit counters .Channel N_Port Xmit Rec 40 BCs BCs can be different across ports Number of BCs on receiver port F_Port Xmit Rec 10 BCs F_Port Xmit Rec 10 BCs Number of BCs on receiver port CU N_Port Xmit Rec 5 BCs Figure 3-4 Buffer credit determination The sender monitors the receiver's ability to accept frames by managing a buffer credit count value of the receiver port. frame transmission to the associated port is temporarily suspended.R_RDY If the buffer credit count is zero (0). This flow control mechanism prevents data overruns. Buffer credit calculation based on full frame size To calculate the number of buffer credits needed to fully utilize an FC link with an average FICON frame size of 2112 bytes. a 10 km FC link running at 8 Gbps for would need 40 buffer credits. Channel N_Port Xmit Rec BC counter 7 Frame 3 Frame 2 Frame 1 10 Buffers Xmit Rec BC counter CU N_Port 5 Buffers 5 Xmit Rec Figure 3-5 Buffer credit counters An R_RDY signal is returned for every received frame when a receive port buffer is made available. The buffer credit count value is decremented by the transmitter when a frame is sent (see Figure 3-5).

or with a very large block size such as 4 x 4 KB or 12 x 4 KB blocks per I/O. After a FICON switched fabric has been customized to support FICON cascaded Directors. During initialization. High integrity helps to ensure that any changes to the data streams are always detected. the FICON channel will query the fabric to determine that it supports high integrity. the CUP must be configured and the FDR parameter in SYS1. and that the data frames are delivered to the correct endpoint (FICON channel or control unit port). Cyclical Redundancy Checking (CRC) and Longitudinal Redundancy Checking (LRC) are bit patterns added to the data streams to allow the detection of any bit changes. reporting. isolation. as shown in Figure 3-7 on page 42. then the channel will complete the initialization process. If an accidental cable swap occurs. High integrity fabric support Another important benefit of FICON support of cascaded Directors is its ability to provide high integrity data paths. which will cause data frames to be delivered to the wrong endpoint. FICON frame size information can be tracked using the FICON Director Activity Report in RMF. For FICON channels. The high integrity function is an integral component of the FICON architecture when configuring FICON channel paths through a cascaded fabric. the Director will check that its Inter-Switch Links (ISLs) are attached to the correct Director before they are made operational. High integrity features are used within the FICON channel and the cascaded FICON Directors to ensure the detection and reporting of any miscabling actions occurring during operational use.It is important to note that not all applications utilize the same frame size. FICON Director technical description 41 . End-to-end data integrity is designed to be maintained through the cascaded FICON Directors. and recovery. The high integrity fabric feature for cascaded FICON Directors protects against miscabling and misdirecting of data streams. For RMF to receive the information. the Director invokes logical path testing. A FICON channel requires that the cascaded FICON Directors support high integrity. If it does. The FICON frame size of 2112 bytes can be achieved through the use of zHPF with large sequential reads/writes.PARMLIB must be set to Yes. allowing the channel to operate with the fabric. and the required World Wide Node Name (WWNN) and Domain IDs have been added in the fabric membership list. Chapter 3.

and isolation. Note: Activating fabric binding is a prerequisite for a cascaded FICON Director topology. Fabric binding Fabric binding is a security feature that enables explicit control over which FICON Directors can be interconnected by preventing non-authorized Directors from merging. The FICON Director that is allowed to connect must be added to the fabric binding database of the other FICON Director. Channel initialization completes. For example. This is done by manually defining the authorized Director in a fabric binding database. and error recovery. miscabling occurs (for example. if a Director has this 42 FICON Planning and Implementation Guide . 5. 6.5 3 FICON Director 1 4 ! FICON Director 6 System z servers oops Storage Devices 2 FICON Director FICON Director 6 Figure 3-7 High integrity fabric feature The checking process proceeds in this way: 1. Insistent Domain IDs Insistent Domain ID support restricts a switch address (Domain ID) from being automatically changed when a duplicate switch address is added to the fabric. Any I/O requests to the invalid route are discarded until the error is corrected. Data is protected. At some later time. The Director port enters “invalid attachment state” and notifies state change back to System z. reporting. either accidentally or intentionally. The System z server invokes the channel logical path testing. 4. The fabric binding database contains the World Wide Node Name (WWNN) and Domain ID of the connecting FICON Director. Overt operator action is required to change a switch address. 2. cables are swapped at a patch panel). Insistent Domain ID support also prohibits the use of dynamic Domain IDs to ensure that predictable Domain IDs are enforced within the fabric. 3.

Zoning Zoning is a method used in the FICON Director to enable or disable communication between different attached devices.feature enabled. A zone consists of a group of ports or WWNs. Chapter 3. default zone members are automatically removed whenever that member is added to an active zone. which restricts port connectivity based on port number WWN zoning is typically used for open systems connections. Nodes that are not already defined to a zone are automatically put into the default zone. FICON Director technical description 43 . which permits connectivity between attached nodes based on WWN Port zoning. such as: In-band management via a FICON Channel (using the CUP) Out-of-band management via the following: – IP-based client/server management software – Simple network management protocol (SNMP) – Trivial file transfer protocol (TFTP) to load firmware – Command-line Interface (CLI) via Telnet – GUI provided by the Director vendor or other vendors CLI via the serial interface (RS232) Call home via a modem connection (for notification purposes) Figure 3-8 on page 44 shows the various interfaces that can be used to manage the FICON Director. Conversely. This means that the SAN infrastructure can be shared by both protocols. Connectivity is permitted only between connections to the Director that are in the same zone. Port zones should be used for FICON. Management capabilities FICON Directors can be managed using different communication methods. to create a commonly managed infrastructure. the new Director is segmented into a separate fabric and user data will not flow. There are two types of zoning: Name zoning. Protocol intermix An intermix of FICON and Fibre Channel Protocol (FCP) is supported by FICON Directors at the port level (each port can run either FICON or FCP). and a new Director is connected to it (through an ISL) without the Domain ID feature enabled.

CLI (via Telnet). 44 FICON Planning and Implementation Guide . as well as simulate failure and recovery actions. The Cisco MDS 9500 Series encompasses the Cisco MDS 9513. and TFTP Vendor GUI or Web-based Interface SA for z/OS and RMF (using the CUP) Out-of-band via Ethernet (IP) CLI In-band via the CUP FICON Channel Call Home Serial Interface Figure 3-8 FICON management interfaces 3. availability. and performance) are identified and resolved before the FICON Directors are qualified for System z use. all errors must be detected and properly handled by the attached FICON Directors. New York. The SAN b-type family consists of the IBM SAN768B and the IBM SAN384B. The recovery tests check for the correct detection of a fault or error condition by the attached FICON Directors. In this section. For a test to be classified successful..2 Qualified FICON Directors The purpose of IBM qualification testing for FICON Directors is to ensure that any issues related to interoperability with System z servers (such as functionality. Table 3-1 on page 45 lists the IBM and Cisco FICON Directors and some of their key capabilities. IBM proprietary software and microcode utility test suites stress the FICON Directors to ensure they can perform at levels beyond typical customer usage. In addition. Cisco MDS 9509. and the Cisco MDS 9506. S. The test facilities provide the capability to configure a fully functional FICON environment to drive the various components and protocols of the FICON Directors. we focus on five of the qualified FICON Directors that support a link data rate of 8 Gbps: two from the IBM System Storage SAN b-type family and three from the Cisco MDS 9500 Series. and ensure that the recovery adheres to System z architecture rules. Qualification testing is carried out in the IBM Vendor Solutions Connectivity (VSC) Lab in Poughkeepsie. SNMP. data integrity. U.

10 2. However. 4. thereby eliminating oversubscription. any ingress/egress traffic moving through port groups on the same blade does not cross the backplane (local switching). 4. FCP FICON. 32-port. 2. FCP Protocol intermix Yes Yes Yes Yes Yes Number of ports 16 to 384 16 to 192 12 to 528 12 to 336 12 to 192 Port speed (Gbps) 1. 4.Table 3-1 Product IBM and Cisco FICON Directors IBM Model 2499-384 2499-192 2054-E11 2054-E07 2054-E04 Supported protocols FICON. FCP FICON. 4. Chapter 3.5 : 1. 8 ISL speed (Gbps) 2. 4. 2. However. Four slots are utilized for two control processor (CP) blades and two core switching blades. FICON Director technical description 45 .536 Tbps of chassis bandwidth based on 192 ports running at 8 Gbps (link data rate). The SAN768B can switch up to 384 ports simultaneously. The eight remaining slots can be used for various blade types.5 : 1. or 48-port 8 Gbps blades Six-port 10 Gbps blade for dark fiber connections (such as ISLs) An encryption blade Routing/extension blade. The SAN384B provides 1. 2. 4. Table 3-2 on page 46 through Table 3-6 on page 47 list the component options available with the SAN768B and SAN384B FICON Directors. Four slots are utilized for two control processor (CP) blades and two core switching blades. 10 2. 4. 4.2. The chassis has 12 slots. 8. The SAN768B utilizes a vertical chassis and measures 14U tall. If all 384 ports are running at 8 Gbps and that traffic travels over the backplane.com/systems/storage/san/index. the oversubscription ratio is 1. 8 1. 8. go to: http://www-03. The chassis has 8 slots. thereby eliminating oversubscription. any ingress/egress traffic moving through port groups on the same blade does not cross the backplane (local switching). 8. the oversubscription ratio is 1. If all 192 ports are running at 8 Gbps and that traffic travels over the backplane.1 IBM System Storage SAN b-type family components The SAN768B and the SAN384B are built on a common platform. 4. FCP FICON. 8. The SAN384B can switch up to 192 ports simultaneously. FCP FICON. 2. 10 2. 2. The four remaining slots can be used for various blade types. if using FCIP A fabric application blade A future FCoE/CEE blade The SAN768B provides 3. The blade slots in both platforms can be filled with: 16-port. 8 1. 10 IBM System Storage SAN768B IBM System Storage SAN384B Cisco MDS 9513 Cisco MDS 9509 Cisco MDS 9506 To find other System z qualified FICON Directors. 10 2. The SAN384B utilizes a horizontal chassis and measures 8U plus a 1U tall exhaust shelf.072 Tbps chassis bandwidth based on 384 ports running at 8 Gbps (link data rate). 4. 8 1.ibm. 8.html 3. 8 1.

Table 3-2 lists the available blade options and the associated feature codes. Table 3-4 SAN768B and SAN384B licensed and miscellaneous options Licensed and miscellaneous options SAN 768B Inter-Chassis Cable Kit SAN768B Pair of Upgrade Power Supplies Feature code 7870 7880 46 FICON Planning and Implementation Guide . All ports must be populated with SFPs (each with a minimum of 1 Gbps and a maximum of 8 Gbps). Table 3-3 SAN768B and SAN384B transceiver options Transceiver options a 4 Gbps Short Wave SFP Transceiver 4 Gbps WSW 8-Pack SFP Transceiver 4 Gbps 10 km Long Wave SFP Transceiver 4 Gbps 10 km LW 8-Pack SFP Transceiver 4 Gbps 4 km Long Wave SFP Transceiver 4 Gbps 4 km LW SFP Transceiver 8 Pack 4 Gbps 30 km ELW SFP Transceiver 10 Gbps FC SW XFP Transceiver 10 Gbps FC LW XFP Transceiver 1 GbE Copper SFP Transceiver 8 Gbps Short Wave SFP Transceiver 8 Gbps SW SFP Transceiver 8 Pack 8 Gbps 10 km Long Wave SFP Transceiver 8 Gbps 10 km LW SFP Transceiver 8-Pack 8 Gbps 25 km ELW SFP Transceiver a. Table 3-3 lists the available transceiver options and the associated feature codes. Maximum of 48 eight-packs or 255 singles Feature code 2401 2408 2411 2418 2441 2448 2480 2510 2520 2550 2801 2808 2821 2828 2881 Table 3-4 lists the available licensed and miscellaneous options and associated feature codes. Table 3-2 SAN768B and SAN384B blade options Blade options FC 8 Gbps 16-portsa FC 8 Gbps 32-portsa FC 8 Gbps 48-portsa (does not support loop devices) FC 10 Gbps 6-ports FC Routing (16 + 2 ports) Feature code 3816 3832 3848 3870 3850 a.

Licensed and miscellaneous options SAN768B Inter-Chassis License SAN768B FICON w/CUP Activation FCIP/FC High-Performance Extension FICON Accelerator Integrated Routing Feature code 7885 7886 7887 7888 7889 Table 3-5 lists the available fiber optic options and associated feature codes. from ports on storage devices all the way to the FICON channel of the System z server. “Basic functions of a FICON Director” on page 37.2. Chapter 3. DCFM features easy-to-use administrative tools that streamline and automate repetitive tasks. IBM System Storage Data Center Fabric Manager IBM System Storage Data Center Fabric Manager (DCFM) is a comprehensive network management application that enables end-to-end management of data center fabrics. It can be used to configure System Storage SAN b-type family Directors. DCFM is designed for unified management of data center fabrics. Table 3-5 SAN768B and SAN384B fiber optic cable options Fiber optic cable options a Fiber Cable LC/LC 5 meter 50 µm multimode Fiber Cable LC/LC 25 meter 50 µm multimode Fiber Cable LC/LC 5 meter 50 µm multmode 4-Pack Fiber Cable LC/LC 25 meter 50 µm multimode 4-Pack Fiber Cable LC/LC 31 meter 9 µm single mode 4-Pack Fiber Cable LC/LC 31 meter 9 µm single mode a.1. FICON Director technical description 47 . Table 3-6 SAN768B and SAN384B rack mounting options Rack mounting options First SAN768B installed in a 2109-C36 rack Second SAB768B installed in a 2109-C36 rack Standalone mode Feature code 9281 9282 9284 3. Maximum of 96 four-packs or 255 singles Feature code 5305 5325 5405 5425 5444 5731 Table 3-6 lists the available rack mounting options and associated feature codes.4.2 IBM System Storage SAN b-type family functions This section discusses platform-specific functions that are available on System Storage SAN b-type family Directors for use in FICON environments. The platform-specific functions complement the functions that are described in 3.

TI zoning has two modes of operation: Failover enabled With failover enabled. traffic stops and will not be rerouted to another path even if other paths are available outside the TI zone. true traffic isolation is achieved. TI zones are a logical AND with classic zoning (port or WWNN zoning). If all paths in the TI zone fail. host-to-target traffic. When two System Storage SAN b-type family Directors are interconnected by ICLs. Traffic will only take the path specified by the TI zone. The ICL ports on are internally managed as E_Ports and use proprietary connectors instead of traditional SFPs. each chassis still requires a unique Domain ID and is managed as a separate Director. It can be used. For this reason. then traffic will rerouted to another TI zone. The following features are part of the Adaptive Networking suite: Traffic Isolation Routing QoS Ingress Rate Limiting QoS SID/DID Traffic Prioritization Traffic Isolation Routing The Traffic Isolation Routing feature allows you to control the flow of interswitch traffic by creating a dedicated path for traffic flowing from a specific set of source ports (N_Ports). If all paths in the TI zone fail. the Adaptive Networking features can maximize fabric behavior and provide necessary bandwidth for high priority. ICL ports in the core blades are used to interconnect the two Directors. Note that TI zoning does not replace port or WWNN zoning. 48 FICON Planning and Implementation Guide . traffic will take the path specified by the TI zone regardless of other paths that are available. to dedicate an ISL to high priority. Traffic isolation is implemented using a Traffic Isolation zone (TI zone). Even under the worst congestion conditions. for example. Traffic Isolation Traffic Isolation (TI) zoning allows you to direct traffic to certain paths in cascaded FICON Director environments. Failover disabled With failover disabled.DCFM provides multiprotocol networking support for: Fibre Channel Fiber Connectivity (FICON) Fibre Channel over IP (FCIP) Fibre Channel Routing (FCR) Internet SCSI (iSCSI) Inter-Chassis Link An Inter-Chassis Link (ICL) is a licensed feature used to interconnect two System Storage SAN b-type family Directors. mission-critical applications and connections. Adaptive Networking Adaptive Networking is a suite of tools and capabilities that enable you to ensure optimized behavior in the SAN. ISLs (E_Ports) must be included in the TI zone. potentially increasing the number of usable ports.

Ingress rate limiting feature Ingress rate limiting is a licensed feature that requires the Adaptive Networking license.and data-intensive applications Up to eight ISLs can be combined into a trunk.Quality of Service feature Quality of Service (QoS) is a feature in Adaptive Networking advanced fabric services that enables users to set the desired QoS for applications running on both physical and virtual servers. FICON Director technical description 49 . maximum availability. Fabrics can automatically and continuously adapt to changing network dynamics to ensure that critical applications always receive high priority. Top Talkers in Advanced Performance Monitoring feature The Top Talkers feature is part of the licensed Advanced Performance Monitoring feature. A High QoS setting gives critical applications the required bandwidth priority. because each LISL is dedicated to traffic for a single fabric. It can restrict the speed of traffic from a particular device to the Director port. Virtual ISL connection When a base switch is connected to another base switch. providing up to 68 Gbps data transfers (with 8 Gbps ISLs). Lossless Dynamic Load Sharing Lossless Dynamic Load Sharing (DLS) allows you to rebalance trunk port paths without causing input/output (I/O) failures in cases where the end devices require in-order delivery (IOD) of frames. You can use DLS on trunks connecting switches to perform the following functions: Eliminate dropped frames and I/O failures by rebalancing the paths going over the ISLs whenever there is a fabric event that might result in sub-optimal utilization of the ISLs Eliminate the delay caused by establishing a new path when a topology change occurs Pause ingress traffic (by not returning credits) Chapter 3. Frame-level trunking can: Optimize link usage by evenly distributing traffic across all ISLs at the frame level Maintain in-order delivery to ensure data reliability Help ensure reliability and availability even if a link in the trunk fails Optimize fabric-wide performance and load balancing with Dynamic Path Selection (DPS) Simplify management by reducing the number of ISLs required Provide a high-performance solution for network. Top Talkers can identify the ports that consume the most bandwidth and that information can be used to assist you with configuring the SID/DID pairs with certain QoS attributes so they get proper priority. The LISL isolates traffic from multiple fabrics. When logical switches with the same fabric ID (FID) are configured to use the XISL (such as the logical switches #1 and #2). It measures and ranks top bandwidth-consuming traffic in real time based on SID/DID pairs. and optimal performance in the event of network congestion. Ingress rate limiting can also be used to reduce existing congestion in the network or proactively avoid congestion. the base switches automatically create a Logical ISL (LISL) within the XISL. Frame-level trunking Frame-level trunking automatically distributes data flows over multiple physical Inter-Switch Link (ISL) connections and logically combines them into a trunk to provide full bandwidth utilization while reducing congestion. a Virtual ISL (XISL) connection is created.

Ports are numbered from 0 through 15 from bottom to top on the left set of ports. or traffic type (FC. control. Only port-based numbering is permitted in the default switch. each of which is assigned to a separate logical fabric.2. The base switch can only contain ISLs that are shared by other logical switches. A logical switch can have any user-defined FID between 0-127. Although not supported for FICON.0 and above. The default switch is FID = 128 and it is where all ports are placed when virtual fabrics are enabled. A logical switch and logical fabric are independent entities that are managed just like a physical switch or fabric. the default switch is usually not used for FICON. business group. Table 3-7 Logical switch types Switch type Base switch Description Not supported for FICON. the base switch can be used for other logical switches. Table 3-8 shows the port numbering scheme for each blade type. Table 3-8 Port numbering scheme Blade type 16-port 32-port Blade port numbering Ports are numbered from 0 through 15 from bottom to top. Logical switches support both port-based and zero-based port numbering. customer. Vitalization supports the following use cases: Isolating traffic types Pooling resources for independent fabrics Isolating different application tiers and device management Facilitating backbone edge routing Increasing FICON consolidation and scalability Connect logical and “unaware” switches to form logical fabrics Isolate and manage by application. As a result. Virtual fabrics are enabled by default for all switches shipped from the factory with FOS 6. FICON) Table 3-7 lists the logical switch types supported by the Virtual Fabrics feature. and 16 through 31 from bottom to top on the right set of ports. Ports are identified by both the slot number in which the blade is located in the chassis and the port number on the blade (slot number/port number).Change the existing path to a more optimal path Wait for sufficient time for frames already received to be transmitted Resume traffic after a failure Virtual fabrics or logically partition a physical SAN The Virtual Fabrics feature enables a physical storage area network (SAN) switch to be partitioned into multiple logical switches. The logical switch is the most common switch type used for FICON. and management isolation for logical switches and logical fabrics using a standards-based implementation of the ANSI standard for virtual fabrics. Zero-based port numbering is not supported. 50 FICON Planning and Implementation Guide . The Virtual Fabrics feature provides data. Default switch Logical switch Port numbering A port number is a number assigned to an external port to give it a unique identifier to the FICON Director.

Table 3-9 lists some of the functions and their associated hardware. QoS.0c and above FOS 6.software.html Chapter 3. Some forward planning.0c Inter-Chassis Link (ICL) Traffic Isolation (TI) Virtual Fabric (logical partition) Quality of Service (QoS) Adaptive Network TopTalker.html IBM System Storage SAN384B Web page: http://www-03.Blade type 48-port 6-port Blade port numbering Ports are numbered from 0 through 23 from bottom to top on the left set of ports. GC52-1333 Brocade Fabric OS Administrator’s Guide.0. Any port can be assigned to any logical or default switch. you can perform a migration without making any changes to IOCP or HCD.0. Ingress Rate Limiting SAN768B or SAN384B SAN768B or SAN384B SAN768B or SAN384B SAN768B or SAN384B SAN768B or SAN384B For more details regarding the features and functions provided by the System Storage SAN b-type family. and firmware prerequisites that are discussed in this section for the SAN768B and SAN384B. ports can have any number the user defines. however.0. 53-1001185 IBM System Storage SAN768B Web page: http://www-03. refer to the following: IBM System Storage SAN768B Director Installation. Service and User's Guide.ibm.2 FOS 6. software. and firmware prerequisites Function Data Center Fabric Manager Hardware prerequisites All supported Directors Software and firmware prerequisites Refer to: ftp://ftp. For each FICON Director to be migrated. a logical switch with its Domain ID and port numbers identical to those in the old FICON Director can be created.ibm.com/systems/storage/san/b-type/san768b/index. software. and 24 through 47 from bottom to top on the right set of ports. FICON Director technical description 51 . can simplify the task of numbering the ports and minimize service impacts.0c and above FOS 6. Ports are numbered from 0 through 5 from bottom to top. Migration The combination of logical switches and zero-based addressing make for easy migration from IBM System Storage SAN b-type and m-type FICON Directors to a SAN768B or SAN384B platform.ibm. When zero-based addressing is used.com/systems/storage/san/b-type/san384b/index. GA32-0574 IBM System Storage SAN384B Director Installation.com/common/ ssi/pm/sp/n/tsd03064usen/TSD03064U SEN. Table 3-9 Hardware.1 FOS 6.PDF FOS 6. Service and User's Guide. Therefore.

The Cisco MDS 9506 has 4 slots for various modules and can simultaneously switch up to 192 ports.1 or later) Feature code 2404 2412 2424 2448 2450 2824 2844 2848 Table 3-11 lists the available transceiver options and associated feature codes. storage services.2. The Cisco MDS 9506 has dual Supervisor modules and measures 7U tall. and multi-service modules 12-port. The Cisco MDS 9513 has 11 slots for various modules and can switch up to 528 ports simultaneously. Table 3-10 through Table 3-14 on page 53 list the component options available with the Cisco MDS 9500 Series FICON Directors. when all 24-ports module are used simultaneously at 8 Gbps.2 or later) 24 8 Gbps FC Switching Module (requires OS 4.1 or later) 48 port 8 Gbps FC Switching Module (requires OS 4. Table 3-11 Cisco MDS 9500 Series transceiver Transceiver options FC 10 Gbps SRX2 Transceiver FC Ethernet 10 Gbps SRX2 Transceiver FC 10 Gbps 10 km LWX2 SC Transceiver FC 10Gbps 40 km ERX2 Transceiver Feature code 5030 5032 5040 5050 52 FICON Planning and Implementation Guide . On a 24-port switching.1.1 or later) 4/44 port 8 Gbps Host Switching Module (requires OS 4. 48-port. the oversubscription factor could be as much as 4 : 1.2. and 4/44-port 8 Gbps switching modules 4-port 10 Gbps switching module When all 48 ports on a 48-port switching module are used simultaneously at 8 Gbps. The module slots in all three platforms can be filled with: IP services. the oversubscription factor could be as much as 2 : 1. The Cisco MDS 9509 has 7 slots for various modules and can switch up to 336 ports simultaneously. 24-port. Table 3-10 lists the available line card options and associated feature codes.1. The Cisco MDS 9500 Series utilizes a vertical chassis.3. The Cisco MDS 9513 and MDS 9509 with their dual Supervisor-2 modules measure 14U tall. Table 3-10 Cisco MDS 9500 Series line card options Line card options 4 port 10 Gbps FC Switch Module 12 port 4 Gbps FC Switch Module 24 port 4 Gbps FC Switch Module 48 port 4 Gbps FC Switch Module 18/4 Multservice Module (requires OS 3. and 48-port 4 Gbps switching modules 24-port.1.3 Cisco MDS 9500 Series The Cisco MDS 9500 Series is a modular multi-layer platform with common switching and service modules.

1 or later) Tri-Rate SW SFP Transceiver Tri-Rate LW SFP Transceiver Gigabit Ethernet Copper SFP (feature code 2450 is required) Feature code 5434 5444 5454 5830 5834 5850 5854 5210 5220 5250 Table 3-12 lists the available hardware and software packages and associated feature codes. Table 3-12 Cisco MDS 9500 Series hardware and software packages Hardware and software packages MDS 9500 Enterprise Package MDS 9500 Fabric Manager Server Package MDS 9500 Mainframe Server Package Feature code 7021 7026 7036 Table 3-13 lists the available fiber optic cable options and associated feature codes.2.4 Pack 25 meter 50 µm LC/LC Fiber Cable .4 Pack FC 4 Gbps 4 km LW SFP Transceiver 4 Pack FC 4 Gbps 10 km LW SFP Transceiver . FICON Director technical description 53 .1.4.1.1 or later) FC 8 Gbps 10 km LW SFP+ Transceiver (requires OS 4.1 or later) FC 8 Gbps 10 km LW SFP+ Transceiver .1.1.1.1 or later) FC 8 Gbps SW SFP+ Transceiver .Transceiver options FC 4 Gbps SW SFP Transceiver . The platform-specific functions complement the functions that are described in 3.4 Pack FC 8 Gbps SW SFP+ Transceiver (requires OS 4. Table 3-13 Cisco MDS 9500 Series fiber optic cable options Fiber optic cable options 5 meter 50 µm LC/LC Fiber Cable (multimode) 25 meter 50 µm LC/LC Fiber Cable (multimode) 5 meter 50 µm LC/LC Fiber Cable .4 Pack Feature code 5605 5625 5642 5643 Table 3-14 lists the available rack mounting options and associated feature codes. Table 3-14 Cisco MDS 9500 Series rack mounting options Rack mounting options Plant Install 9513 in 2109-C36 rack Field Merge 9513 in 2109-C36 rack Feature code 9543 9544 3. “Basic functions of a FICON Director” on page 37. Chapter 3.4 Pack (requires OS 4.4 Pack (requires OS 4.4 Functions of the Cisco MDS 9500 Series This section discusses platform-specific functions that are available on the Cisco MDS 9500 Series of Directors for use in FICON environments.

so a failure of a switching module cannot bring down the PortChannel link. It also enables accumulation and analysis of historical performance data. Secure File Transfer Protocol (SFTP). it is referred to as a TE port. Summary View displays a summary of E_Ports (Inter-Switch Links). Simple Network Management Protocol Version 3 (SNMPv3). It provides a graphical user interface (GUI) that displays real-time views of your network fabric and lets you manage the configuration of Cisco MDS 9500 Series FICON Directors. is a feature specific to switches in the Cisco MDS 9500 Series. 2. F_Ports (fabric ports). role-based access control. If trunk mode is enabled on an E_Port and that port becomes operational. over the same physical link. Comprehensive security framework The Comprehensive security framework supports RADIUS authentication. and provides access to statistics and configuration information for a single switch. each Director in switched point-to-point or cascaded FICON Director topology. Fibre Channel Security Protocol (FC-SP). Secure Shell Protocol (SSH). VSANs. Device View displays a graphic representation of the switch configuration.Fabric Manager The Fabric Manager is a set of network management tools that supports secure Simple Network Management Protocol version 3 (SNMPv3) and earlier versions. Detailed traffic analysis is also provided by capturing data with SNMP. The trunking feature includes the following: Trunking configurations are applicable to E_Ports. port security. The captured data is compiled into various graphs and charts that can be viewed with any Web browser. the trunking protocol ensures seamless operation as an E_Port 54 FICON Planning and Implementation Guide . which can be a valuable complement to RMF. also known as VSAN trunking. is a separate fabric). Trunk E port (TE port) Trunking. Device Manager The Device Manager presents two views of a Director: 1. It adds the ability for Fabric Manager to manage multiple fabrics simultaneously (note that in Fibre Channel terms. using enhanced ISL (EISL) frame format. The trunk-allowed VSANs configured for TE ports are used by the trunking protocol to determine the allowed-active VSANs in which frames can be received or transmitted. PortChannel (ISL Trunking) PortChannels refer to the aggregation of multiple physical interfaces into one logical interface to provide higher aggregated bandwidth. fabric binding and VSAN-based access control. hardware-enforced zoning. load balancing. PortChannels can connect to interfaces across switching modules. and N_Ports (attached hosts and storage) on the Director. The Cisco Fabric Manager applications are: Fabric Manager Server Device Manager Fabric Manager Server Fabric Manager Server is a separately licensed feature that extends the capabilities of the basic Fabric Manager. and link redundancy. If a trunking enabled E_Port is connected to a third-party switch. Trunking enables interconnect ports to transmit and receive frames in more than one VSAN.

or high priority). one VSAN with 10 ports can span 10 different line cards. Port numbers do not change based on TE ports. isolated environments within a single physical fabric for secure sharing of physical infrastructure and enhanced FICON intermix support. and the port addresses can be swapped. and VSAN-administrator roles can be set up to only allow configuration and management of specific VSANs. FICON Control Unit Port. and eight of those can be used for FICON. medium. chassis-wide unique port numbers should be reserved for TE ports. Port numbering A range of 250 port numbers is available for you to assign to all the ports on a Director. Adding ports to a VSAN is a nondisruptive process. You can also assign specific VSANs to certain cascaded links. and security profile. a network administrator role can be set up to allow configuration of all platform-specific capabilities. You can have more than 250 physical ports assigned and the excess ports do not have port numbers in the default numbering scheme. VSAN-based access control reduces SAN disruptions by localizing the effects of user errors to the VSANs for which the user has administrative privileges. Domain ID. FICON Director technical description 55 . name server. The maximum number of ports for a VSAN is 255 due to FICON addressing rules. or you can assign duplicate port numbers if they are not used in the same FICON VSAN. Quality of service (QoS) offers the following: Provides a relative bandwidth guarantee to application traffic Controls the latency experienced by application traffic Prioritizes one application over another (for example. Each VSAN has its own set of fabric services (fabric server. port numbers are the same as port addresses. By default. VSANs enable complete traffic isolation and are maintained throughout the FICON fabric. For example. Virtual Storage Area Network Virtual Storage Area Network (SAN) technology is used for hardware-enforced. Fabric Shortest Path First (FSPF) routing. VSAN-Based Access Control This feature enables customers to define roles where the scope of the roles is limited to certain VSANs. The following rules apply to FICON port numbers: Supervisor modules do not have port number assignments. Because TE ports appear in multiple VSANs. It adds a layer of security where only administrators can configure switches within specified VSANs. For example. prioritizing transactional traffic over bulk traffic) through bandwidth and latency differentiation You can apply QoS to ensure that Fibre Channel data traffic for your latency-sensitive applications receive higher priority over throughput-intensive applications such as data warehousing. operating mode. IP address.Quality of Service (QoS) The QoS feature in Cisco SAN-OS allows for data traffic to be classified into distinct levels for service differentiation ( low. The Cisco MDS 9500 Series can have up to 256 unique VSANs. VSANs can span line cards and are dynamic in size. Chapter 3. You can have ports without a port number assigned if they are not in a FICON VSAN. and so on).

the relevant PortChannel configuration is applied to the physical port. For each FICON Director to be migrated. For more details regarding the features and functions provided by the Cisco MDS 9500 Series. go to: http://www. Each FCIP tunnel must be explicitly associated with a FICON port number.html Cisco MDS 9506 for IBM System Storage http://www-03.com/systems/storage/san/ctype/9509/index.cisco. Migration The combination of VSANs and FICON port numbers make for easy migration from previous-generation FICON Directors to a Cisco MDS 9500 infrastructure. refer to the following: Cisco MDS 9513 for IBM System Storage http://www-03. If the port numbers are not assigned for PortChannels or for FCIP tunnels.com/en/US/partner/products/ps5990/tsd_products_support_series_ home.html For additional Cisco MDS 9500 Series-related information. Therefore. you can perform a migration without making any changes to IOCP or HCD.ibm. When the port number for a physical PortChannel becomes uninstalled. a FICON VSAN with its Domain ID and FICON port numbers identical to those in the old FICON Director can be created. then the associated ports will not come up.com/systems/storage/san/ctype/9506/index.html Cisco MDS 9509 for IBM System Storage http://www-03.com/systems/storage/san/ctype/9513/index.ibm.html 56 FICON Planning and Implementation Guide .ibm.Each PortChannel must be explicitly associated with a FICON port number.

2009. 2005. It also gives a solid base for configuring the components of the solution during the implementation phase. 57 .Part 2 Part 2 Planning the FICON environment This part provides an overview of the items that you need to consider when planning the FICON environment. © Copyright IBM Corp. It provides a structured approach for planning and designing an end-to-end FICON infrastructure solution. 2006. All rights reserved.

58 FICON Planning and Implementation Guide .

4

Chapter 4.

Planning the FICON environment
When you are building a communication infrastructure with the highest quality of service for a FICON environment, proper planning is crucial. This chapter provides you with a structured approach for planning and designing an end-to-end FICON infrastructure solution. It gives a solid base for configuring the components of the solution during the implementation phase.

© Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

59

4.1 Structured approach for planning
Having a structured approach for planning is key to reaching your desired goal when you are designing any end-to-end solution. The approach used in this chapter offers a framework that is based on best practices and industry standards. The topics we cover follow a logical sequence that aids in the decision-making process for planning a FICON environment. The decision points are listed here and described in greater detail in the following sections: Step 1 - Documentation Step 2 - Requirements Step 3 - Context Step 4 - Topologies and supported distances Step 5 - Convergence Step 6 - Management Step 7 - Virtualization and availability Step 8 - Performance Step 9 - Prerequisites and interoperability Step 10 - Physical connectivity If you are not familiar with Fibre Channel and FICON architectures and terminology, then before proceeding we strongly recommend that you read Chapter 1, “Introduction to FICON” on page 3; Chapter 2, “System z FICON technical description” on page 13; and Chapter 3, “FICON Director technical description” on page 33. In addition, to be successful in designing a FICON solution, you must be well versed in the technologies and products that support typical FICON solutions. Although this book provides some guidance, it is not meant to provide in-depth technology and product knowledge. Therefore, we recommend that you also attend the appropriate training and education offered by IBM as described at the following URL: http://www-304.ibm.com/jct03001c/services/learning/ites.wss/zz/en?pageType=page&c= a0001742

Step 1 - Documentation
Current and accurate documentation (planning, design, installation, and final) is essential for a successful implementation and later operations of the FICON environment. However, because most information is not always readily available during the various phases of the planning process, there will be situations where it will require iterations of the process. To avoid delays during the planning and implementation process, ensure that the various types of documentation needed are always well-defined, complete, and current.

Step 2 - Requirements
Planning for any end-to-end solution is a process of iterations. It starts with gathering requirements and defining the desired outcome of the solution. After all relevant data has been collected, you will need to review and evaluate it, as well as outline the objectives of the solution. If you are migrating from an existing environment, you must also consider the current configuration, including the functions and features that are implemented and how they are being used.

60

FICON Planning and Implementation Guide

Step 3 - Context
The output of the requirements-gathering phase should provide a solid base for mapping to one or more of the following scenarios: Migrating from ESCON to FICON Moving to an 8 Gbps environment (FICON Express8) Migrating from a single site to a multi-site environment Implementing a new FICON environment

Step 4 - Topologies and supported distances
To fulfill the requirements of the chosen solution (or scenario), you must select the most suitable topology. The total end-to-end distance (from a channel to a control unit) is a main factor in determining the correct topology. The following topologies are supported: Point-to-point Switched point-to-point Cascaded FICON Directors For extended distances, you must also evaluate the IBM qualified solutions depending on your requirements and the supported distances of the FICON connections. Note that IBM owns a strict qualification process for giving assurance of an extended distance solution based on WDM technology.

Step 5 - Convergence
The I/O communication infrastructure (FICON fabric) should provide the ability to attain high availability and continuous operations in a predictable manner, with centralized management and scalability for future changes. This means that the switching platform must be robust enough to deliver such capability, as well as be ready to deploy new enhancements that improve throughput, security, and virtualization. Therefore, the FICON Director is the reference platform for FICON fabrics. Factors that influence the selection of the FICON Director platform are: Intermix of the FICON and Fibre Channel Protocol (FCP), which enables you to consolidate the I/O infrastructure, maximize the utilization of the bandwidth, and consolidate the fabric management. Security, high integrity, and zoning are integral parts of the entire FICON infrastructure. Other relevant factors for the FICON fabric, such as power consumption, cooling, and space inside a data center, should also be considered.

Step 6 - Management
The selection of the systems management platforms and their integration with the operation support environment is very important to the continuous operations of the FICON infrastructure. The possibilities for FICON management include the following: Command-line Interface Element management Fabric management Storage management initiative specification System z management for FICON Directors
Chapter 4. Planning the FICON environment

61

Step 7 - Virtualization and availability
Because the efficiency and effectiveness of a data center is determined by the levels of virtualization and availability, the complete infrastructure must be considered. All aspects of virtualization and physical connectivity for the FICON channel, Director, and control unit should be evaluated.

Step 8 - Performance
The selection of the right technology is determined by satisfying traffic patterns, further segmentation, and performance of the end-to-end solution. It is very important to understand the fan-in, fan-out, and oversubscription ratios for the entire FICON environment. Traffic management gives you the possibility to proactively manage any congestion inside the fabric and minimize latency. There are a number of design evaluation tools that can aid in this process.

Step 9 - Prerequisites and interoperability
Checking the compliance of all FICON components in the infrastructure with the prerequisites of the planned solution is key to ensuring that your implementation will proceed without delays and avoid interoperability problems. The FICON components include the System z FICON features and control units, as well as the FICON Directors and Wavelength Division Multiplexor (WDMs) platforms, if applicable to the selected topology.

Step 10 - Physical connectivity
The physical transport media has to be selected according to the installed or planned features. Choosing a qualified cabling solution based on industry standards for Data Center Cabling (US TIA-942/Europe EN50173-5) and best practices in accordance with System z qualified IBM Facilities Cabling Services will provide a secure and highly scalable consolidated physical layer for the complete data center. In the subsequent sections we describe considerations and recommendations for each of the identified decision points in more detail.

4.2 Documentation
Planning, design, installation, and final documentation is a requirement for a successful implementation and later operations of the FICON environment. Based on best practices and requirements we recommend the following types of documentation:

Planning documentation
Current environment Requirements Available features and functions Decision points

Design documentation
High level Detailed level

Installation documentation
Detailed implementation plan

62

FICON Planning and Implementation Guide

Final documentation
Technical Operations Disaster recovery plan Change procedure The number of documents that you must create depends on the complexity of the solution and the required level of detail. Creating and maintaining documentation throughout the planning, design, and implementation phases of your FICON infrastructure is an important part of running the environment without unplanned outages. For the most part, the implementation, operations, and support (troubleshooting) responsibilities are owned by different groups. From a security viewpoint, there is usually a a clear boundary between the responsibilities. Changes have to be monitored and audited to ensure that due diligence is observed; for example, execution must be strictly separated from auditing. And finally, you have to consider the planned and unplanned turnover of professionals at the workplace. To reduce the risk of disruption to your FICON infrastructure, keep your documentation as current and accurate as possible and store a copy in a secure place, preferably in a different location in case of a disaster.

4.3 Requirements
It is important to clearly understand and accurately document all requirements. Such documentation will help you throughout the planning and design process. After the requirements are collected and documented, each one will have to be carefully evaluated. For existing environments (FICON or ESCON), it is also important to identify all equipment that is currently installed and how it is being used. This means physical and logical inventories should be carried out. The goal of the physical inventory is to identify and verify what is installed in your FICON or ESCON environment. Through onsite visual inspections and documentation, all ESCON or FICON channels, Directors and ports, control units, and cabling should be identified. Although the overall goal of the physical inventory is to identify and verify what you have physically installed, the goal of the logical inventory is to understand the functions that are being used and how they are defined. Determine whether the functions must be replaced by a FICON function. During the planning and design process, the requirements will be mapped to features, functions, topology, and technologies, which in turn will determine an approximate design of your FICON environment. This does not mean you will only have one solution, but rather multiple alternatives. The alternatives will have to be assessed using all the steps in this chapter.

4.4 Context
To define the context of your environment, you must take into consideration the existing and planned components (System z, FICON Director, and control units) that were determined and
Chapter 4. Planning the FICON environment

63

documented in the requirements gathering phase. You will then have to define your transport layer strategy dependent on the upper layer requirements for high availability, disaster recovery, and business continuity. This all has to be synchronized with the communication layer requirements of your applications. After all relevant data is reviewed and analyzed, the output (alternatives) can be used to map to one or more of the following scenarios: Migrating from ESCON to FICON (changing the transport layer protocol) Moving to a high bandwidth environment (implementing FICON Express8) Migrating from a single site to a multi-site cascaded topology (building a highly available multi-site solution) Implementing a new FICON environment (building a completely new FICON environment) The following sections provide addition information for each scenario.

4.4.1 Migrating from ESCON to FICON
This scenario discusses the considerations for migration from an ESCON environment to a FICON environment. We concentrate on DASD migration, because it is the most commonly connected device type. FICON is addressing technical limitations of ESCON in bandwidth, distances, topology, channel and device addressing and performance. For the right number of FICON channels you need: The I/O operations per second and their characteristics The required bandwidth ESCON channels are planned with 50% utilization or less and depend on block size. With higher utilization, the traffic will be slower. Another limit is the I/O rate, which is typically less than 1000 I/O per second. The ESCON data droop is caused by the single CCW transmit at a time, and then waiting for the notification. We do not recommend planning over the actual limits, because that can cause performance degradation. It is also important to have enough channels to keep the number of concurrent I/Os low. You also have to determine the peak I/O and MB/s values for each subsystem. This reflects current workload and planned growth. We recommend that you use analysis tools over a longer time period to obtain realistic trends. For native tape, throughput is very important. Aggregating multiple ESCON tape channels onto a single FICON channel reduces the System z infrastructure. This can be reached due to high head-to-tape transfer rates, capacity, and accessibility of the cartridge. Using FICON for virtual tapes also reduces the number of required CUs, back-end tape drives, and channel paths. When planning the migration from an existing ESCON configuration to a FICON configuration, consider the following rules: Configure at least two channel paths to a logical control unit for high availability. (More channel paths may be required, depending on throughput requirements.) A logical control unit or device cannot be defined to be accessed more than once from the same channel path.

64

FICON Planning and Implementation Guide

“FICON Director technical description” on page 33. Chapter 3. and efficiency of your System z and storage environment. but the access is to different CUADDs (different logical control units) within the physical control unit.number of ports and bandwidth Fabric growth . the 8 Gbps features only auto-negotiate to speeds of either 8 Gbps.A physical control unit that has multiple logical control units (specified by the CUADD parameter in IOCP/HCD) may be accessed more than once from the same FICON (FC) channel path.end-to-end management Backup . scalability. We also recommend that you review the capabilities of the existing cabling infrastructure. or 2 Gbps. Chapter 4. We recommend that you do not mix low speed and high speed devices on the same 8 Gbps channel. Also. the advantages of using FICON include: More concurrent I/Os to the same control unit Concurrent I/Os with other control units Link data rate (200 Mbps for ESCON versus up to 8 Gbps for FICON) Unrepeated distance from channel (3 km for ESCON and 10 km for FICON) 4. With the use of a FICON Director you can utilize the physical channels even more when shared with multiple connections to different CUs at the same time. Even if there are the same number of paths from the operating system image to the disk subsystem in both ESCON and FICON configurations.4. With an 8 Gbps FICON environment. while improving performance. explains the differences in proprietary implementations for each product line.I/O throughput and scalability Services . you need to address the following requirements: Storage growth . At long distances. The physical layer up to 4 Gbps link data rate had different limitations as compared to 8 Gbps and 10 Gbps. Configure the channel paths according to the quantity of resources available in the FICON channel and control unit. Planning the FICON environment 65 . Building a high-performance infrastructure that provides the flexibility to deploy 8 Gbps as needed simplifies the fabric management and scales with the growth of applications and data.bandwidth and cabling infrastructure Server virtualization .scalability and efficiency Compatibility .2 Moving to a high bandwidth environment (FICON Express8) Moving to 8 Gbps FICON improves the performance. Some benefits of using 8 Gbps FICON include: The ISL oversubscription ratio is halved by upgrading from 4 Gbps to 8 Gbps ISLs. 4 Gbps. 8 Gbps compared to lower speeds optimizes the number of dark fiber or WDM links.end-to-end management and lifetime There are two System z qualified FICON Director platforms that support 8 Gbps: the IBM System Storage b-type family and the Cisco MDS 9500 Series.bandwidth Operational flexibility .number of ISLs and scalability Performance . It is not possible to aggregate two or more ESCON channel paths that access the same logical control unit into only one FICON channel path.

TSD00070USEN Cisco MDS 9513 for IBM System Storage. SG24-6116 IBM System Storage SAN768B. Cascaded FICON Directors allow for FICON channels to connect a System z server to another. and cooling Scalability and flexibility . the greater distances you can go without impacting the response times. FICON Directors typically have more buffer credits per port than the server and the disk.For more detailed information. as well as the potential for fiber infrastructure cost savings for extended storage networks. Solutions such as Geographically Dispersed Parallel Sysplex™ (GDPS®) can benefit from the reduced intersite configuration complexity that FICON support of cascaded Directors provides. TSD01754USEN 4.4 Implementing a new FICON environment When you are defining a new FICON environment. The faster the channel speeds between sites. TSD00069USEN Cisco MDS 9509 for IBM System Storage. “FICON Director technical description” on page 33 Implementing an IBM/Brocade SAN with 8 Gbps Directors and Switches. power. similar server or to peripheral devices (such as disk. With higher speed this factor can even multiply. and storage devices. The following factors affect the performance of a cascaded FICON Director configuration: Number of ISLs and the routing policy Number of FICON channels routed across the ISLs ISL bandwidth ISL traffic management I/O workload characteristics Distances between the components Buffer-to-buffer credits FICON support of cascaded Directors (sometimes referred to as “cascaded switching” or “two-switch cascaded fabric”) is for single-vendor fabrics only. we recommend that you take advantage of the latest FICON functions and features that support high performance and high availability.3 Migrating from a single site to a multi-site environment Migrating from a single site to a multi-site configuration usually means going from a point-to-point or switched point-to-point topology to a cascaded topology. The more buffer credits you have. or printer) via two FICON Directors. refer to: Chapter 2. Cascaded support is important for disaster recovery and business continuity solutions.floor space.4. the better the intersite throughput. 4. It also reduces the number of FICON channels. Director ports. In cascaded environments. the cabling for both intersite and intrasite can be simplified. The sharing of links between the two sites reduces the number of physical channels between sites and the number of Director ports. It can provide high availability connectivity. This decreases the management complexity. For planning this scenario you must also consider criteria such as: Facility requirements . tape library. or the tape subsystems. TSD03037USEN Cisco MDS 9506 for IBM System Storage.4.ability to add ports and bandwidth dynamically 66 FICON Planning and Implementation Guide . “System z FICON technical description” on page 13 Chapter 3.

06 1.1 1. 4.85 2. The link loss budget is derived from the channel insertion loss budget defined by the FC-PI-4 standard (Revision 8.4 a.Performance .maximum distances and link loss budget FC-PI-4 Fiber core (light source) 1 Gbps Distance in meters (in feet) Link loss budget in dB 2 Gbps Distance in meters (in feet) Link loss budget in dB 4 Gbps Distance in meters (in feet) Link loss budget in dB 8 Gbps Distance in meters (in feet) Link loss budget in dB 10 Gbps ISLa Distance in meters (in feet) Link loss budget in dB 9 µm SM (10 km LX laser) 9 µm SM ( 4 km LX laser) 50 µm MMb (SX laser) 50 µm MMc (SX laser) 62.58 2.04 300 (984) 82 (269) 33 (108) 2.I/O throughput and traffic patterns Availability .8 10000 (32736) 7.6 3.00). Inter-Switch Link (ISL) between two FICON Directors b.8 4000 (13200) 4.8 4000 (13200) 4.2 fiber c.4 10000 (32736) 6.8 NA NA NA NA 860 (2822) 500 (1640) 300 (984) 4. OM2: 50/125 µm multimode fiber with a bandwidth of 500 MHzkm at 850 nm and 500 MHz-km at 1300 nm in accordance with IEC 60793-2-10 Type A1a.31 380 (1247) 150 (492) 70 (230) 2.8 10000 (32736) 6.78 1.62 500 (1640) 300 (984) 150 (492) 3. bandwidth.62 2.4 (FC-PI-4) standard and used by System z FICON features.88 150 (492) 50 (164) 21 (69) 2. Planning the FICON environment 67 . using best practices and industry standards that also align with vendor recommendations.5 µm MMd (SX laser) 10000 (32736) 7. and logical paths End-to-end management Cabling infrastructure Distances between the components Control unit and device capabilities Interoperability of all components The remaining sections in this chapter help you define a baseline for your FICON infrastructure.5/125 µm multimode fiber with a minimum overfilled launch bandwidth of 200 MHzkm at 850 nm and 500 MHz-km at 1300 nm in accordance with IEC 60793-2-10 Type A1b fiber Chapter 4. OM3: 50/125 µm laser optimized multimode fiber with a minimum overfilled launch bandwidth of 1500 MHz-km at 850nm as well as an effective laser launch bandwidth of 2000 MHz-km at 850 nm in accordance with IEC 60793-2-10 Type A1a.8 10000 (32736) 7.3 3. Table 4-1 identifies cabling types and link data rates that are supported in the FICON environment. OM1: 62.0 4000 (13200) 4. Table 4-1 Fiber optic cabling for FICON .redundant ports. including their allowable maximum distances and link loss budget.Physical Interface .68 2.5 Topologies and supported distances The unrepeated distance supported by System z FICON features is dependent upon: Transceiver type Fiber optic cabling type Port speed Mode conditioning patch (MCP) cables Cabling specifications are defined by the Fibre Channel .1 fiber d.0 2.

4.Note: IBM does not support a mix of 50 µm and 62.43 miles) for 1.43 miles) for 1 Gbps LX links with RPQ 8P2263 12 km (7. 2. the maximum unrepeated distance of a FICON LX link using a single mode fiber optic cable is: 10 km (6.5 µm fiber optic cabling in the same physical link. 4. one FICON Director increases the maximum supported distance of a FICON LX channel path using single mode fiber optic cables to the following distances: 20 km (12. “System z FICON feature support” on page 25 for details about the FICON features available on the System z10 and System z9 servers. 8 Gbps LX links 30 km (18.3 Cascaded FICON Directors A FICON channel path may include a maximum of two cascaded FICON Directors.5.64 miles) for 1 Gbps LX links with RPQ 8P2263 22 km (13.1 Point-to-point As illustrated in Figure 4-1. 4. The use of extended distance longwave transceivers on the Inter-switch links (ISL) between the 68 FICON Planning and Implementation Guide .2.5. 8 Gbps LX links 20 km (12.2 Switched point-to-point As illustrated in Figure 4-2.46 miles) for 2 Gbps LX links with RPQ 8P2263 FICON link 9 µm fiber FICON CU System z 10 km >10 km (RPQ 8P2263) Figure 4-1 FICON LX point-to-point maximum unrepeated distance 4.67 miles) for 2 Gbps LX links with RPQ 8P2263 FICON Director FICON CU FICON link 9 µm fiber FICON links 9 µm fiber FICON CU FICON CU System z 10 km >10 km (RPQ 8P2263) 10 km Figure 4-2 FICON LX switched point-to-point maximum unrepeated distance 4. 2.5.2 miles) for 1. Refer to 2.

Each link in the channel path can be either long wavelength (LX) or short wavelength (SX). the maximum supported distance of a FICON LX channel path using single mode fiber optic cables is: 30 km (18.86 miles) for 1 Gbps LX links with RPQ 8P2263 32 km (19.FICON Directors may be required.64 miles) for 1. As illustrated in Figure 4-3. 4. allowing the channel path to be made up of an intermix of link types as illustrated in Figure 4-4. Each ISL requires one fiber trunk (two fibers) between the FICON Directors.88 miles) for 2 Gbps LX links with RPQ 8P2263 FICON Director FICON Director FICON CU FICON link 9 µm fiber ISL 9 µm fiber FICON links 9 µm fiber FICON CU FICON CU System z 10 km >10 km (RPQ8P2263) 10 km 10 km Figure 4-3 FICON LX cascaded FICON Directors maximum unrepeated distance A FICON channel path through one or two FICON Directors consists of multiple optical fiber links. and using the example of 10 km ISL links between FICON Directors. 2. The transceiver types (LX or SX) at each end of a given link must match. Planning the FICON environment 69 . Site 2 Site 1 FICON Director SM LX FICON CU LX SX LX LX LX MM SM SX FICON CU LX FICON CU 10 km SM M LX SX SM System z LX LX SX LX LX SX LX SM LX FICON CU Figure 4-4 FICON Director media intermix M MM SM SX FICON CU LX FICON CU FICON Director Chapter 4. 8 Gbps LX links 40 km (24. This is possible because the FICON Director performs an optical-to-electrical-to-optical conversion (OEO) of the channel path as it passes through the Director.

as illustrated in Figure 4-5. DWDM technology also provides increased flexibility because multiple links and protocol types can be transported over a single dark fiber trunk.4 Extended distance The most important element for an extended distance topology is latency. The same end-to-end distance is also available between the FICON Director port and the control unit port. Note: The maximum distance of the fiber optic link for a FICON Express8 LX feature to the first hop is limited to 10 km. M M SX FICON CU SM LX FICON CU LX SX SX LX FICON Director MM SM SX FICON CU LX FICON CU 70 FICON Planning and Implementation Guide .5. For all FICON features using repeaters. the FICON SX and LX implementation supports a maximum distance of 100 km between two sites. as illustrated in Figure 4-6 on page 71. Using Dense Wavelength Division Multiplexing (DWDM) technology. However. because of the time it takes for the light to travel over distance The opto-electro-opto converter (or the WDM equipment). Site 2 LX FICON CU DWDM Site 1 LX km 100 SM SM LX FICON CU LX SX SX LX FICON Director MM DWDM SX FICON CU LX FICON CU SX SX SM SX System z Figure 4-5 FICON point-to-point and switched point-to-point maximum extended distance By combining FICON cascaded Director technology with DWDM technology. the overall end-to-end distance between the FICON channel and control unit is System z-qualified up to 100 km only. if used This has to be compensated for by the buffer credit technology used inside the interface logic of the FICON Director or the channels at the System z server and the control unit.4. the FICON implementation supports a maximum distance of 100 km between two sites. The repeated distance for a FICON channel is System z-qualified to a maximum of 100 km (62 miles). the end-to-end distance between the FICON channel and the FICON Director port can be up to 100 km. Latency is usually caused by the following: The length of the fiber.

Planning the FICON environment 71 .Site 2 Site 1 FICON Director SM LX FICON CU LX SX LX LX LX LX SX SX SX SX MM SX FICON CU LX FICON CU LX SM 10 0 DW DM SM SX SX km DW DM M M SX SX System z LX LX LX LX SX SX SM LX FICON CU SX SX LX SX LX MM SM SX FICON CU LX FICON CU FICON Director Figure 4-6 Cascaded FICON Directors maximum extended distance For more detailed information. you can skip this section and go to 4. GA23-0367 For the list of IBM qualified extended distance FICON Director solutions. see: http://www-03. high-bandwidth applications. cooling.7.ibm. This approach also simplifies management and reduces power consumption. and Open Systems Adapters).html For the list of IBM qualified extended distance WDM solutions. and simplify hardware and cabling. Other relevant factors to consider are: High density port count (small form factor pluggable ports) Best energy effectiveness devices (per transferred byte) High availability (a redundant architecture of all components) High speed (switching capability . refer to: Planning for Fiber Optic Links (ESCON. Chapter 4. FICON. and space requirements inside a data center. see: http://resourcelink.com/systems/storage/san/index.backplane throughput) If you selected a point-to-point topology in the previous step.com 4. Coupling Links. “Management” on page 75.6 Convergence Some System z data center fabrics may require the use of a fully integrated multiprotocol infrastructure that supports mixed I/O and traffic types for simultaneous storage connectivity. A converged infrastructure enables you to consolidate multiple transport layers in a single physical interconnect. This consolidation provides the flexibility to build virtual server and storage environments supporting low-latency.ibm.

The management differences between the two protocols are not relevant unless you want to control the scope of the switching through zoning or connectivity control. cooling. System z servers accessing remote storage. it queries the Name Server for the zoning information.1 Intermix fabric Following are reasons why you might move to an intermix FICON and FCP environment: Because of an open systems SAN that is delivering reliable. FCP configurations support multiple fabrics and allow seven hops between source and target. For example. FCP devices use name server zoning as a way to provide fabric-wide connection control. FCP communications are name-centric. and they use the Fibre Channel Name Server to determine device communication. The FICON Director is required to support hardware enforcement of the connectivity control. FICON devices do not query because the allowable port and device relationships have been previously defined. FICON abstracts the concept of the port by creating an object known as the port address. Hardware-based remote disk mirroring. An association is then made between the port address and the port number. although other techniques complementary to zoning (such as port. It is also important to understand the implications of FICON port addressing versus port numbering in FCP. and other operational tasks. partitioning. and fabric binding) are used. this is more restrictive than the zoning information used by open systems devices. Director. any device in the fabric can access the ANSI standard FC management server by logging into the fabric. 72 FICON Planning and Implementation Guide . Possible scenarios include the following: FICON and open systems environments with a common storage network. Binding helps alleviate the security concerns that are experienced in intermix installations. To consolidate switching and cabling infrastructure. you must block the transfer of any and all frames from a FICON Director port to all SAN connected ports and vice versa. high-performance connectivity. binding. and the management requirements of the target environment. In intermix environments. and other design considerations. This concept facilitates the ability to perform FICON port swaps for maintenance operations without the need to regenerate the host configuration. discovery-oriented. these fabrics often have available ports on their SAN Directors. FCP port-to-port connectivity has traditionally been enforced via zoning. Important technical considerations that you must take into account include fabric management techniques such as zoning. When an FCP device attaches to the fabric. The port address abstraction is not in the Fibre Channel architecture and is foreign to FCP. The PDCM is a vector-defined addressing system that establishes which addresses are allowed to communicate with each other. which enables more effective maintenance. First you must consider the management techniques used in each protocol.4.6. because with FCP. and fabric-assigned. power consumption. Different System z server LPARs are accessing local or remote storage via FICON or via FCP. If you are implementing intermix. FICON devices can use the Prohibit Dynamic Connectivity Mask (PDCM) to provide Director-wide connection control. Linux on System z to access local storage or remote storage.

confidentiality and disclosure. Securing access to the management tools. by using: – Physical separation – Electronic separation – Logical separation The use of Intrusion Detection and Incident Response – Automatic event detection and management – Event forensics – Incident procedure to exercise due diligence Chapter 4. by using: Securing Long-Distance SAN. SG24-7545 4. SG24-7544 Implementing an IBM/Cisco SAN. and integrity and alteration of data. However. SG24-7543 IBM System Storage/Brocade Multiprotocol Routing: An Introduction and Implementation. by using: – – – – – – – – – – – – – – – – – – HTTPS Secure Shell (SSH) Secure Socket Layer (SSL) Secure Copy (SCP) LDAP-MS Active Directory® Role-Based Access Control (RBAC) Password policies IP Filters (firewall) Passive FTP RADIUS server Zoning Fabric Configuration Server (FCS) Fabric Configuration Server (FCS) policy Switch Connection Control (SCC) policy Device Connection Control (DCC) policy CHAP and DH-CHAP IPsec P-EAP and MS-CHAP Securing the fabric.For more detailed information about these topics. Therefore. Planning the FICON environment 73 . SG24-6116 IBM/Cisco Multiprotocol Routing: An Introduction and Implementation. any SAN security strategy should also include the following: Restrict administrator privileges Isolate sensitive environments Audit SAN activities All-level documentation You need to consider the following topics when planning a secure fabric. it is important to remember that the majority of security threats are from insiders.2 Fabric security Best security practices typically address three main objectives: availability and destruction.6. refer to: Implementing an IBM/Brocade SAN with 8 Gbps Directors and Switches.

then strict fabric binding is mandatory (even with one FICON Director in the fabric). The default zone automatically puts nodes attached to a fabric that are not already in a zone into a single zone and allow any-to-any connectivity. 4. ensure that you use values within the FICON Director’s range. the FICON channel queries the Director to determine that it supports high-integrity architecture. Conversely. and that the data frames are delivered to the right endpoint.3 High integrity Data integrity ensures that any changes to the data streams are always detected. Domain ID and switch id have to be unique. A strict fabric-wide consistency is necessary for FICON Director binding. We recommend that you set the switch ID (in IOCP or HCD) to the same value as the domain ID of the FICON Director. default zone members are automatically removed whenever that member is added to an active zone. When defining the switch IDs in the IOCP definitions. There are two types of zoning: WWN zoning and port zoning. and to control traffic paths. allowing the channel to operate with the fabric. for example. you ensure that the maximum number of zone members is not exceeded and does not inadvertently allow unintended connectivity. High-integrity fabric architecture support includes fabric binding and insistent Domain ID support. 4. 74 FICON Planning and Implementation Guide . A FICON Director provides the following fabric. you can use the encryption capabilities of System z. Although this is not supported with FICON. even if the port zone includes all the ports in the fabric. which is defined to the FICON Director at installation time. If it does. This policy prevents unauthorized Directors from joining a fabric. If you use 2-byte addressing on the System z. This policy binds device ports to Director ports. This simplifies the configuration and reduces confusion by having a common designation across all definitions.6. to restrict confidential data to specific servers.The use of Fabric-Based Encryption This technology enables you to encrypt specific or all disk data to prevent accidental data leaks when disk drives are replaced or disk arrays refreshed. and port binding features: Fabric binding is a security method for restricting Directors within a multiple-Director fabric. A FICON channel requires the support of high-integrity architecture when configured to operate in a cascaded Director fabric.4 Zoning Zoning is a method used in the Director to restrict communication. The best practice recommendation is to disable the default zone and create port zones. the channel completes the initialization process.6. By disabling the default zone and defining port zones. Zoning is used. to separate FICON devices from each other. Port binding is a security method for restricting server or storage devices that connect to particular Director ports. Director. During initialization. and to protect enterprises from data theft. Director binding is a security method for restricting devices that connect to a particular Director.

even if LUN masking is used. FICON prohibits dynamic connectivity mask (PDCM) controls whether or not communication between a pair of ports in the Director is prohibited or allowed. Each level has a specific mission to fulfill. more zones can provide additional security. only the FICON ports are put in the port zone. stable. Use single initiator zoning with separate zones for tape and disk traffic if an HBA is used for both traffic types. Define port zones for FICON and World Wide Name (WWN) zones for FCP. this does not provide enough control over the entire environment. Use port WWN identification for all zoning configuration. Planning the FICON environment 75 . Disk and tape mirroring protocols are FCP even when the front end is FICON.7 Management We recommend that you manage the FICON fabric at various levels. Use frame-based hardware enforcement. and easy to manage: Implement zoning. The only exception to this is the loss of synchronization link incident. use standard open systems best practices for creating WWN zones for the FCP traffic to ensure that the SAN is secure. Adding WWN zoning on top of that adds an unnecessary layer of complexity. Persistently disable all unused ports to increase security. Ports can easily be added to port zones even if there is nothing attached to the port. Unfortunately. The most simple is offered by the CLI. A serial connection may not be possible if the FICON Director is physically located in a restricted or remote area. the most restrictive rules are automatically applied. Changing the configuration of a FICON Director without a Telnet connection requires a serial connection. That node can be moved anywhere in the fabric and it will remain in the same zone.7. because every time a new channel card or control unit interface is added to the fabric. Chapter 4. In intermixed environments. We recommend that you separate this traffic from the FICON traffic either with WWN zoning or virtual fabrics. Port zoning is what should be used for FICON. the WWN zoning would have to be modified. To create a solid base for a SAN. Limit the aliases and names to allow maximum scaling. 4. Disable access between devices inside the default zone. If there are any differences in the restrictions set up and PDCM.WWN zoning permits connectivity between attached nodes based on WWN. Link incidents for FICON ports are reported only to registered FICON listener ports. To operate the FICON infrastructure you need to automate the whole process. In purely FICON environments. The IOCP effectively does what WWN zoning does in open systems environments. WWN zoning is used in open systems environments and is not used for FICON. 4. All management levels create together robust systems management and interconnections with data center management. you put all ports in a single port zone. In more complex fabrics. Port zoning limits port connectivity based on port number.1 Command-line interface The command-line interface (CLI) uses role-based access control (RBAC) to manage access for operations of the FICON Director.

then they must have the same subnet mask and have a common IP addressing range. Web Tools provides an extensive set of features that enable you to quickly and easily perform key administrative tasks: Setting up IP addresses. or workstation from any location with IP connectivity (Analogy to Cisco MDS Device Manager). and RADIUS Enabling Ports on Demand Routing setup Configuring links and ISL trunks 4.7.2 Element management Web Tools enables an intuitive and easy-to-use interface which helps to monitor and manage FICON Directors.) DCFM Enterprise includes following: High scalability of managed Director ports Multiprotocol support Status and connectivity visualization End-to-end performance monitoring Fault management Troubleshooting and diagnostics Call home Offline support analysis Configuration management Security management FICON environment configuration and management with analysis and diagnostic capabilities Virtual fabric management 4. SMI-S is the only standard that addresses manageability from the perspective of a device in a storage array. license keys. 4. For recommended tasks executed from the CLI.Because all FICON Directors ship with the same default IP address. desktop PC. If all FICON Directors will be connected to the same management network. refer to platform-specific documentation.3 Fabric management A client-server based application for managing the whole multi-fabric environment is called the Data Center Fabric Manager (DCFM).4 Storage management initiative specification The storage management initiative specification (SMI-S) defines the interface that allows storage management systems to manage and monitor storage area network (SAN) resources.7. You will also need a sufficient number of IP addresses for each FICON Director dependent on the platform.7. 76 FICON Planning and Implementation Guide . to an application running on a host. You can perform tasks by using a Java™-capable Web browser from your laptop. (Analogy to Cisco MDS Fabric Manager Server or Cisco Data Center Network Manager. only one can be connected to the management network at a time. all the way through the data path. Director names. and Simple Network Management Protocol Performing firmware upgrades and controlling boot parameters Maintaining RBAC.

It enables you to receive RMF Monitor I reports for reporting on FICON Director activity statistics in the SMF Record Type 74 Subtype 7. Best practice is to have two or more CHPIDs with access to the CUP for backup. It enables you to use the Dynamic Channel Path Management function of Workload Manager and IRD (DCM) with FICON. This prevents local commands from interfering with host-based commands by starting serial access to Director parameters.ibm. For more details about IBM Storage Management Platform. Resource Monitoring Facility (RMF) and the CUP can provide activity statistics and resource utilization. FICON Directors have an embedded port FE in the Control Processor for the CUP function. RMF Monitor I performs long-term data collection for resource utilization. Too much activity to the CUP can cause missing interrupts. It enables you to get Service Information Messages (SIMs) to the z/OS console for FICON device hardware failures. Control unit port The control unit port (CUP) used by System z provides in-band management of the FICON Director. FICON management server mode must be enabled on the Director to enable CUP management features.7.5 System z management for FICON Directors Combined. We make the following recommendations: Only one RMF should attempt to access the CUP at any given time. On a 256-port Director. refer to the following URL: http://www. such as System Automation z/OS for I/O-Ops. If processing times are increased.The SMI-S does not require any modification or upgrade to deployed fabrics when it is deployed. It can be a demon running directly on the Director platform or a separate host application. this logical FE address overlaps the physical port FE address. IBM Tivoli System Automation for z/OS and the Control unit port (CUP) of the FICON Director have the capability to control and automate the entire I/O communication. The implementation is vendor-specific.storage. Planning the FICON environment 77 . it is possible to inadvertently time out the channel because the CUP is busier than normal. You also have to configure the Missing Interrupt Handler process timeout. When Chapter 4. The advantage of using an end-to-end storage management platform is the possibility of management automation of the whole storage network including the FICON fabric. You should use the CUP for the following reasons: It enables you to perform in-band management from the Hardware Management Console (HMC). Each fabric device needs its own specific agent. A FICON Director that supports CUP can be controlled by one or more host-based management tools. It enables you to use the management tools used with ESCON. A mode register controls the behavior of the FICON Director according to CUP and other management interfaces. which determines when the channel will time out waiting for an expected response from the CUP.com 4.

However. refer to FICON Director-specific documentation. For more details. Virtualizing the FICON components enables a flexible and dynamic resource sharing solution. I/O operations changes paths by allowing you to control channels. I/O operations It provides a single point of control for managing connectivity in your active I/O configurations. and provides a connection from a focal point processor to a target processor. control units can only allocate a limited number of logical paths in relation to the number of logical paths that FICON channels can request.1 System z The use of FICON channels gives you the ability to define shared channels that can request logical paths. consider virtualizing your FICON components. Tivoli System Automation for z/OS performs the following major operations: System operations It monitors and controls system operations.8 Virtualization and availability To achieve the most effective FICON infrastructure. which can involve using dynamic switching. In this case. physical port addresses FE and FF cannot be used. The Prohibit dynamic connectivity mask and port names are for ports 0 through 253. Processor operations It monitors and controls processor hardware operations. It takes an active role in detecting unusual I/O conditions and allows you to view and change paths between a processor and an input/output device. behavior. ports. use physical port addresses FE and FF for port swapping and for intermix ports. and automate a wide range of system elements spanning both the hardware and software resources of your infrastructure. For information about the CUP license. In the following sections we discuss the virtualization capabilities of the FICON components and what you need to take into account when considering high availability for your FICON infrastructure. and input/output devices.using the CUP. such as the FICON Director. you can refer to: IBM Tivoli System Automation for z/OS Enterprise Automation. 4. Directors. and subsystems. and addressing rules. SG24-7308 4.8. IBM Tivoli System Automation for z/OS With this management platform you can monitor. In configurations 78 FICON Planning and Implementation Guide . we always recommend that you have at least two physical connections to multiple FICON components to avoid single points of failures and ensure high availability. control units. applications. In addition. Ports 254 and 255 are reserved. The following applies to FICON Directors with at least 256 ports in use: The FICON Director is seen by the System z server as a 256-port Director. control. on 256 port boxes. You can do this through an operator console or API.

Multiple Channel Subsystem The design of System z servers offers a considerable increase in processing power. For more information about this topic. thus giving the LPAR a channel image of the shared channel that it can use. memory sizes. REDP-4387 Multiple Image Facility Multiple Image Facility (MIF) allows channels to be shared among multiple LPARs. Create I/O configurations that can exceed the logical path capacity of all or some of the control units in the physical configuration. In support of the larger I/O capability. This capability can be useful or even necessary in several configuration scenarios. Shared channels are configured to an LPAR. Multiple Subchannel Sets Multiple Subchannel Sets (MSS) provides relief for I/O device configurations in large System z environments. a shared channel can reduce hardware requirements without a corresponding reduction in Chapter 4. channels. with a current implementation of two sets. you must manage logical paths to help ensure that the I/O operations take place. and the Logical Channel Subsystem (LCSS) concept is designed to do that. With proper planning. Channel Subsystem Each server has its own Channel Subsystem (CSS). The CSS also provides communication between logical partitions within a physical server using internal channels. as explained in the following sections.where channels request more logical paths than a control unit can allocate. but at the same time provide the capability to selectively establish logical connectivity between control units and channels as needed. Concepts were introduced to facilitate this architectural change and provide relief for the number of supported LPARs. By providing the logical equivalent of multiple physical channels dedicated to multiple LPARs. Planning the FICON environment 79 . and devices available to the server. The CSS enables communication from server memory to peripherals via channel connections. Several components of System z provide the capability for this virtual environment. you can create I/O configuration definitions that allow control units in the configuration to allocate logical paths for every possible request made by channels in either of the following ways: Create a one-to-one correspondence between the logical path capacity of all control units in the physical configuration and the channels attempting to request them. It also increases Parallel Access Volume (PAV) connectivity. Each channel image allows an LPAR to independently access and control the shared channel as though it were a physical channel assigned to the LPAR. and I/O connectivity. Subchannel numbers (including their implied path information to a device) are limited to four hexadecimal digits by hardware and software architectures. The channels in the CSS permit transfer of data between main storage and I/O devices or other servers under the control of a channel program. The CSS allows channel I/O operations to continue independently of other operations within the server. the CSS has been scaled up correspondingly. The solution gives you the ability to have sets of subchannels (addresses). PAV has made this limitation of subchannels a challenge for larger installations. A single disk drive (with PAV) often consumes at least four subchannels. refer to: Multiple Subchannel Sets: An Implementation View. This allows other functions to resume after an I/O operation has been initiated.

You cannot mix shared and unshared channel paths to the same control unit or device. and control unit ports. Channel spanning extends the MIF concept of sharing channels across logical partitions to sharing channels across logical partitions and channel subsystems. Director ports. additional channels can be added for availability. The number of channels is based on the number of concurrent data transfers that the control unit is capable of. This reduction in hardware requirements can apply to physical channels. MIF further improves control unit connection topologies for System z servers with multiple LPARs. For more details see: System z Enterprise Class System Overview. When the channel utilization of shared channels will be greater than unshared channels If you use shared channels to consolidate channel resources. and reduces the complexity of your IOCP input. SC28-6879 IBM System z Connectivity Handbook. the channels can be transparently shared by any or all of the configured logical partitions. Installation can take advantage of MIF performance enhancements offered by: Understanding and utilizing I/O-busy management enhancements Planning for concurrent data transfer Understanding examples of MIF consolidation Understanding and utilizing I/O-busy management enhancements Before you can consolidate channels. Using shared channels reduces the number of channels required. There are some configurations where using an unshared channel is more appropriate. you must consider the channel utilization of all the channels you consolidate. allows for increased channel utilization. The channel utilization of a shared channel will roughly equal the sum of the channel utilizations of each unshared channel that it consolidates. there are some ESCON-capable control units that can only communicate with one LPAR at a time. Note that not all ESCON or FICON configurations benefit from the use of shared channels. When defined that way.I/O connectivity. SG24-5444 80 FICON Planning and Implementation Guide . thereby reducing the number of channels and control unit interfaces required without a corresponding reduction in I/O connectivity. SA22-7832 System z Input/Output Configuration Program User’s Guide for ICP IOCP. you must be aware of the channel requirements of the particular control units you are configuring. MIF enables many LPARs to share a physical channel path. SB10-7037 System z Processor resource/Systems Manager Planning Guide. then you must consider using unshared channels or a different configuration of shared and unshared channels to meet your connectivity needs. If this total channel utilization is capable of decreasing performance. The number of channels needed is independent of the number of LPARs on a system. regardless of the Logical Channel Subsystem to which the logical partition is configured. SB10-7153 System z Support Element Operations Guide. Principles of Operation. SA22-1084 z/Architecture. as explained here: When there are logical path limitations of the control unit Although many ESCON control units can communicate with multiple LPARs at a time using multiple logical paths. depending on the configuration. Although the recommended number of channels satisfies connectivity and performance requirements. MIF allows you to use shared channels when defining shared devices.

On a DS8000 storage unit. A high performance FICON feature provides a significant reduction in channel utilization. IBM HyperPAV is only supported on FICON channel paths. The number of offered CU channel adapters is further increasing the ability to share the storage. you can expose 64 control-unit images (16384 devices) to each System z channel. and can also allow for a reduction in the number of FICON channels required to support a given workload.2 Control unit Consider the following factors concerning the allocation of logical paths by control units: Control units allocate logical paths dynamically on a first come-first served basis. The FICON infrastructure significantly affects the volume of logical path requests to a control unit as follows: Control units can attach to one or more ports on a Director or to additional ports on other Directors. Parallel Access Volume (PAV) enables a single System z server to simultaneously process multiple I/O operations to the same logical volume. To fully access 65280 devices on a storage unit. You can access the devices through a FICON Director to a single storage unit FICON channel. This can increase the number of logical paths that a single channel requests. each logical partition attaching to the same control unit will compete for the control unit’s logical paths. This reduction can allow more I/O input on a single channel. With this method. each storage unit Fibre Channel adapter has four channels. For System z servers. Shared channels require the establishment of a logical path for each channel image corresponding to an active LPAR sharing the channel.4. This is achieved by defining multiple addresses per volume. the Fibre Channel CU channel adapter can operate with fabric or point-to-point topologies. On FICON. which can help to significantly reduce device queue delays (IOSQ time). IBM HyperPAV associates the volumes with either an alias address or a specified base logical volume number. With PAV. No double updates are possible to preserve integrity. I/O configuration definitions for individual control units are not coordinated automatically among the IOCDSs of the different servers. Writes to different domains (a set of tracks the disk controller is working on) are simultaneous as well.8. Each Director port can dynamically connect to many other ports to which channels requesting logical paths are attached. The CU channel adapters maintain compatibility with the existing infrastructure through auto-negotiation. You can further enhance PAV by adding the IBM HyperPAV feature. Planning the FICON environment 81 . Each System z server competes for a control unit’s logical paths. writes to the same domain are serialized. the assignment of addresses to volumes can be automatically managed to help the workload meet its performance objectives and reduce overall queuing. Chapter 4. The CU FICON channels limit the number of devices per channel to 16384. Control units do not manage the allocation of logical paths but instead allow channels to compete for logical paths until all of the control unit’s logical paths are used. you need to connect a minimum of four FICON channels to the storage unit. In a configuration where control units are shared by different servers. reads are simultaneous. which reduces the number of required channels. Each channel has a unique World Wide Port Name (WWPN). However. Control units vary in the number of logical paths they support. FICON support offers a high throughput performance for a single channel. With dynamic PAV.

plug both fiber optic cables into different port cards in the Director. the ISLs needs to be on the same ASIC. GC35-0515 4. SG24-6786 IBM System Storage DS8000 Introduction and Planning Guide. If two paths are defined to attach the CU to the server through the Director. connect the fiber optic cables to ports on different cards. Distribute the CU ports among different port cards. consider where you will connect your FICON channels. based on your availability requirements. For a complete list of features refer to Chapter 3. use the worksheet provided in Appendix B. applications. For frame-based trunking. Distribute the channels among different port cards. 82 FICON Planning and Implementation Guide .8. For more details refer to Chapter 3.3 FICON Director The Virtual Fabrics feature provides data. Before installing the FICON Director. control. connect both fiber optic cables to ports on different cards. Manual settings should only be used with equipment that cannot auto-negotiate or login properly. If two or more ISLs are to be attached between the Directors. Virtualization allows you to: Isolate traffic types. “FICON Director technical description” on page 33 and to platform-specific documentation. “Configuration worksheets” on page 251. to help document the physical layout of your FICON Director. “FICON Director technical description” on page 33 and to platform-specific documentation. refer to: IBM System Storage DS8000 Architecture and Implementation.For more information about these topics. The following topics influence the logical layout of virtual fabrics. Following these rules will ensure that there is always one path available between the server and the CU in case of a defective port card in the Director. FICON cascading only supports an interconnection between FICON Directors of the same platform. For port addressing. but it does not replace them. and management Pool resources Increase FICON consolidation and scalability You have to consider how many virtual fabrics for FICON with FICON Management Server you can create based on vendor limitations. the only interfaces that require manual settings are older 1 Gbps interfaces and long distance ISLs. Distribute the ISLs across different port cards. The Virtual Fabrics option can be added to the existing fabric capabilities. Ports will automatically determine what they are connected to and at what speed they will operate. All standard Director features are supported per virtual fabric. and management isolation for logical Directors and logical fabrics according the ANSI standard for Virtual Fabrics. If two channels are defined to access the same CU. Port characteristics The port addressing mode is platform-specific. Typically.

Plan on having enough ISL bandwidth so that the ISLs do not become congested when the maximum number of streaming tape devices are running. XISL and LISL technology is not supported for FICON at this time. traffic patterns. Director ports. This is vendor-specific.9 Performance A single FICON channel can replace multiple ESCON channels. because that could result in tape traffic termination. A general guideline is to plan for one ISL to four FICON channels that are routed over it. In addition. An understanding of the factors that affect performance must be established before deciding on your target configuration. FICON and FCP traffic should not share the same ISLs. there should always be two ISLs between any two cascaded FICON Directors. and system performance. dependent on the configuration. Tape performance is adversely affected when tapes stop. When logical switches with the same Fabric ID are configured to use the XISL. For Cisco. For information regarding the support of virtual ISLs in FICON environments. resulting in a simpler configuration to manage. This means fewer channels. FICON supports the combined CTC and channel function. and control unit ports. consider spreading those paths across different Directors. Additionally. Planning the FICON environment 83 . which can transport virtual SAN Fabric information between Directors. 4. Use TI zones to direct traffic to different ISLs. For availability. For example. Disk and tape should not share the same ISLs. ESCON-to-FICON channel aggregation can be anywhere between 8:1 and 2:1. Tapes are also more likely to break when stopping and starting. workload characteristics. Virtual ISL When two Directors are connected.If multiple FICON Directors are planned based on the number of required connections and paths. the characteristics of your workload have a direct impact on the performance you will realize in your FICON environment. an ISL connection for two logical switches or extended ISL (XISL) connection for two base switches is created. the Director automatically creates a logical ISL (LISL) within the XISL. The ISL speed must reflect the speed of the channel. the equivalent of virtual ISL is Trunking E-ports. keep the utilization of your FICON channels at or below 50% utilization to maintain optimum response times. Inter-switch link The number of Inter-switch links (ISLs) must be carefully planned based on the existing and planned topology. The following items will require analysis: I/O rates Block sizes Data chaining Read/write ratio Chapter 4. refer to platform-specific documentation. The LISL isolates traffic from multiple fabrics.

0. waiting for access to the data or to reconnect. Disconnect time The channel is not being used for the I/O operation.1 Frame pacing Frame pacing is an FC-4 application data exchange measurement and throttling mechanism. GM13-0702 IBM System z9 I/O and FICON Express4 Channel Performance.For more information about these topics. In the following sections we describe factors that will influence your FICON performance design. and so on). the channel subsystem may not be able to initiate the I/O operation if any path or device busy condition is encountered. Pending time (PEND). by Richard Basener and Catherine Cronin FICON Express2 Channel Performance Version 1. ZSW03059USEN I/O operations Deploying FICON channels demands a further understanding of the I/O operations in your environment. measured by the channel subsystem. Only when using FICON Directors and when the CUP is enabled can RMF provide frame pacing delay information. This is the time that the channel is connected to the control unit. Director. storage port adapter. The I/O Supervisor (IOS) does not issue a start subchannel (SSCH) command to the Channel Subsystem until the current I/O operation to this device ends. thereby freeing the UCB for use by another I/O operation. transferring data for the I/O operation. Frame latency is the average amount of time it takes to deliver a frame from the source port to the destination port. measured by the operating system. is already being used by another I/O request from the same operating system image (UCB busy). ZSW03005USEN IBM System z10 I/O and High Performance FICON for System z Channel Performance. These delays generally result in longer FICON connect time or longer PEND times that show up on the volumes attached to these links. The following terms are used to describe the different phases with respect to the measurements available for determining the duration of an I/O operation. measured by the channel subsystem. After IOS issues the start subchannel command. 84 FICON Planning and Implementation Guide .2x. link distance. It uses buffer credits to provide a flow control mechanism for FICON to assure delivery of data across the FICON fabric. because the control unit is disconnected from the channel. a frame pacing delay occurs.9. After all buffer credits for a port are exhausted. read the following performance papers: Performance Considerations for a Cascaded FICON Director Environment Version 0. I/O supervisor queue time (IOSQ). represented by the UCB. 4. – – – – Channel busy Director port busy Control unit adapter busy Device busy Connect time. The I/O request may be queued in the operating system if the I/O device. Each element that handles the frame contributes to this latency measurement (channel.

we recommend that you upgrade all low speed channels where possible. For read-intensive applications across an ISL (regular transactions). which eliminates handshakes between channel and control unit. the buffer credit value advertised by the E_Port on the target cascaded FICON Director is the major factor that limits performance. A receiver simply advertises buffer credits to the linked transmitter. The buffer credit represents the number of receive buffers supported by a port for receiving frames. The number of buffer credits available for each port on the FICON Director is implementation-dependent. If there are insufficient buffer credits. however. The minimum value of buffer credits is one (1). the processing time at the receiving port. link data rate. Buffer credits are mainly a concern for extended distances. 4. Extended distance FICON enables greater throughput over distance for IBM z/OS global mirror (XRC) using the same distance. This value is used as a controlling parameter in the flow of frames over the link to avoid possible overrun at the receiver. because of the increased number of read commands simultaneously in flight. Planning the FICON environment 85 . For more information about this topic. the buffer credit value advertised by the E_Port on the local FICON Director is the major factor that limits performance. By using the FICON Director activity reports. Depending on the traffic pattern. a poorly designed configuration may consume all available buffer credits and have an impact on performance. or physically separate low and high speed environments. This support allows the channel to remember the last pacing information and use this about subsequent operations to avoid performance degradation at the start of a new I/O operation. In this case the average link utilization goes down. assume you have a FICON 8 Gbps channel attached to two different control units running at lower link rates 4 Gbps and 2 Gbps. There are four implications to consider when planning buffer credit allocation: Ports do not negotiate buffer credits down to the lowest common value. The optimal amount of buffer credits is determined by the distance (frame delivery time). Improved IU pacing (255 IU instead of 16) improves the utilization of the link. it can happen that the low speed device consumes all the available buffer credits of the 8 Gbps link and you are not able to send more packets between the server and FICON Director. refer to “Buffer credits” on page 39. For example. you can gain an understanding of your average read and write frames sizes on a port basis. Extended distance FICON can eliminate the need for channel extenders in 2-site and 3-site z/OS Global Mirror configurations.You must also keep in mind that tape workloads have larger payloads in a FICON frame. while disk workloads might have much smaller. The number of buffer credits is not associated with the performance until high data rates are attempted over long distances. The average payload size for disk is often about 800 to 1500 bytes. there may be a hard limit on the data rate that can be sustained. The exhaustion of buffer credits at any point between an initiator and a target will limit overall performance. and the size of the frames being transmitted.9. Chapter 4. For write-intensive applications across an ISL (tape and disk replication).2 Extended distance FICON The control units that support the Extended distance FICON feature are able to increase the IU pacing count. For this reason.

HyperPAV provides a far more agile alias management algorithm because aliases are dynamically bound to a base for the duration of the I/O for the z/OS image that issued the I/O. Writes to different domains (a set of tracks the disk controller is working on) are simultaneous as well. or are concurrently shared by multiple users 4. Resources that will benefit from multiple allegiance are: Volumes that may have concurrent read operations or a high read-to-write ratio Data sets that have a high read-to-write ratio.9.3 Multiple allegiance Multiple allegiance (MA) is the concurrent operations capability of the IBM disk storage systems that facilitates multiple accesses to or from the same volume with multiple channel path groups and system images. 4. multiple extents on one volume. With PAV.Note. “Extended distance FICON” on page 24. reads are simultaneous. and provide storage capacity and performance improvements. It allows multiple servers to establish concurrent implicit allegiances if there is no possibility that any of the channel programs can alter any data that another may read or write. taking fewer from the 64 K device limitation and leaving more storage for capacity use. – The number of PAV aliases needed might be reduced. For large volumes. Avoid migration to larger volume sizes. Exploit FICON architecture to reduce overhead. to scale I/O rates without the need for additional PAV alias definitions. For more information about this topic. HyperPAV is designed to: Provide a more efficient PAV function. Enable a more dynamic response to changing workloads. No double updates are possible to preserve integrity. This can reduce the PEND time. IBM HyperPAV is only supported on FICON channel paths. however. improve addressing efficiencies. With dynamic PAV. You can further enhance PAV by adding the IBM HyperPAV feature.4 Parallel Access Volume and HyperPAV Parallel Access Volume (PAV) enables a single System z server to simultaneously process multiple I/O operations to the same logical volume. that Extended distance FICON does not extend the achievable physical FICON distances or offer any performance enhancements in a non-z/OS global mirror environment. IBM HyperPAV associates the volumes with either an alias address or a specified base logical volume number. 86 FICON Planning and Implementation Guide . refer to 2.9. However. writes to the same domain are serialized. which can help to significantly reduce device queue delays (IOSQ time). – Dynamic assignment of PAV aliases improves efficiency.1. the assignment of addresses to volumes can be automatically managed to help the workload meet its performance objectives and reduce overall queuing. This is achieved by defining multiple addresses per volume.9. because device busy is not presented to the channel. Simplified management of aliases.

The second reason pertains to FICON multiplexing and the fact that MIDAWs remove some of the data-chaining activity on the link. but MIDAWs do reduce the number of frames and sequences flowing across the link. throughput.com/servers/eserver/zseries/zos/unix/bpxa1ty2. For a homogenous I/O stream of some uniform set of I/Os. In conjunction with Multiple Subchannel Sets (MSS) on System z.com/support/techdocs/atsmastr. for two reasons: One reason pertains to the number of CCWs used. then the average number of I/O operations executing at one time for that LCU during the peak is eight.9. with the advent of MIDAWs. which makes the channel more efficient. you have even more flexibility in device configuration. Multiplexing only affects performance when there are concurrent I/Os on the channel. The most significant performance benefit of MIDAWs is achieved with extended format data sets. Using MIDAWs. If track-level CCWs were “the foundation” for improved FICON performance. By reducing the number of frames. A channel becomes more efficient if the throughput at a particular level of utilization is increased. command chaining was used. Depending on the kind of workload.nsf/WebIndex/TD100311 Use the PAV analysis tool to achieve a better understanding of your workload profile and achieve a more precise split: http://www. For more information about this topic. the I/O rate. The combination of HyperPAV and EAV allows you to significantly reduce the constraint on the 64 K device address limit and in turn increase the amount of addressable storage available on z/OS. Channel efficiency is essentially the relationship of throughput to channel utilization.5 Modified Indirect Data Address Word Using Modified Indirect Data Address Words (MIDAWs) will not cause the bits to move any faster across the FICON link. eight PAV-aliases should be all that is needed to handle the peak I/O rate for the LCU. To span multiple tracks in one channel program. if the average response time is 4 ms and the peak I/O rate is 2000 per second. Planning the FICON environment 87 . Chapter 4. MIDAWs remove the 4 KB boundary restrictions of IDAWs. the alias is returned to the pool for the logical control unit.html 4. Therefore. the z/OS media manager can transfer a whole track using a single CCW. For example. However. The MIDAW facility is a conceptually simple modification to IDAWs. refer to: Disk storage access with DB2 for z/OS.” MIDAWs and IDAWs have basically the same performance characteristics. along with all the other PAV-base addresses in the LCU. and still is used.When an I/O completes. The number of aliases required can be approximated by the peak I/O rates multiplied by the average response time. and channel utilization are all linearly related to each other. the media manager no longer uses data chaining. It then becomes available to subsequent I/Os. and both perform better than data chaining. there is a huge reduction in PAV-alias UCBs with HyperPAV. it takes less time for the FICON channel and the control unit host adapter port to process the channel program.ibm. which in turn affects the number of frames sent across the link. REDP-4187 http://www-03. then MIDAWs represent “the building that rests upon the foundation.ibm.

thereby providing the operating system with the capability to query the state of operation at the control unit when detecting missing interrupts. For more information about this topic. as CHPID type FC. 88 FICON Planning and Implementation Guide .9. zHPF-capable channels and devices support both FICON and zHPF traffic simultaneously. block size can also affect FICON channel performance. The initial implementation of High Performance FICON (zHPF) by the IBM DS8000 is exclusively for I/Os that transfer less than a single track of data. The benefits of zHPF particularly apply to OLTP I/O workloads such as DB2®. or if the channels are faster than the speed at which data can be sequentially read from the disks. Small blocks are usually better for database transaction workloads. However. refer to: How Does the MIDAW Facility Improve the Performance of FICON Channels Using DB2 and Other Workloads?. in addition to eliminating the extended format performance penalty. Missing interrupt problem determination is provided through an interrogate mechanism. There is also dedicated buffer for concurrent sense data now. I/O will flow over the fabric as either FICON or zHPF IUs. MIDAWs affect the performance of update writes in the same manner as they affect reads.6 High Performance FICON Substantial performance improvements can be realized in terms of throughput (MBps) and I/Os per second (number of starts per channel). and control unit log and trace data. except to the extent that the record size affects the number of bytes that fit on a track. MIDAWs help reduce channel utilization and affect the relative performance of different block sizes. and serviceability enhancements. the reason for aborts is provided to control units using an enhanced “purge path” mechanism. VSAM. because the buffer hit ratio for random access tends to be higher with smaller blocks.In addition to reducing channel utilization. FICON continues to use CCWs and zHPF uses TCWs. channel. Significantly more information regarding errors is now provided by the control unit. In other words. availability. There is no change required to the IOCP when using the zHPF architecture. so you have compatibility and coexistence. and zFS which typically transfer small blocks of fixed (4 K) data. There is also no change required to the configuration or design of the fabric. Previously no information was available on missing interrupts and debugging involved attempting to correlate software. Furthermore. The potential exists to achieve some channel aggregation with the adoption of zHPF. PDSE. indirect addressing also has the effect of reducing the utilization of the control unit FICON channel. depending on the application workload. zHPF is transparent to FICON Directors. Physical disk performance is irrelevant if the data permanently resides in cache. using MIDAWs also eliminates the small block performance channel penalty. FICON channels are still defined the same way. The zHPF architecture has built-in reliability. because the system does not have any way of measuring the utilization of the channel. This can result in between a 10% to 30% saving in channel utilization. with or without data chaining. but remember to give careful consideration to meeting all your workload requirements. thereby aiding fault isolation. REDP-4201 4. Realistic production workloads with a mix of data transfer sizes can see between 30% and 70% of FICON I/Os utilizing zHPF. The channel has to process every CCW. thus making it unnecessary to separate sense commands. This effect is hidden. but indirect data addressing is transparent to the channel.

which can cause a congestion on the link even others are not utilized at all. server. and ISL ports that will need to switch with each other within Local Switching groups. we recommend that you use more effective ISL connectivity called frame-based trunking. Combinations of local and non-local traffic flows can occur simultaneously.7 Bandwidth management This section discusses the features and function that can influence the fabric bandwidth.9.For more information about this topic. Because of this. Switching is performed by one or more serial crossbars. The bandwidth maximum for a Director chassis is the crossbar’s bandwidth. This provides highly predictable performance across all ports and an easier design for higher port capacity. server. Fabric Shortest Path First Port speed-weighted Fabric Shortest Path First (FSPF) Fibre Channel routing protocol applies a weighting factor based on link speed to ensure that data traffic uses higher-speed paths between Directors. Switching is simply directing traffic from one port to another. ZSW03058USEN 4. There are three connection allocation possibilities: You can gain a local switching benefit by distributing storage. Planning the FICON environment 89 . Local switching There are two Director architectures: Multi-stage switching with interconnected ASICs The architecture approach uses ASICs to both process and switch data traffic. data flows are automatically distributed over multiple physical Inter-switch link connections. Single-stage switching with one or more crossbars ASICs process data on the port blades. The oversubscription rates can be optimized. it does not assure that all links are equally utilized. enforcing zoning and other protocol-related tasks. to minimize the impact of losing a port blade to a hardware fault or other issue. If more user-facing ports are required. Ensure that array devices are mapped to a local switching group that also has server ports that will require those devices. and ISL ports evenly across port blade ASICs. It gives a basic utilization of multiple ISLs between Directors with the same cost. and ISL ports evenly across Director port blades. two or more ASICs are internally interconnected. By going across port blades and alternating between the high and low ASICs. server. optimizes link usage by evenly distributing traffic across all ISLs at the frame Chapter 4. local switching occurs much more frequently. depending on the design. Frame-based trunking Using frame-based trunking. but they do not switch the traffic. Frame-based trunking combines up to eight ISLs into a single logical trunk. You can group storage. You can perform a slight modification by distributing storage. refer to: High Performance FICON for System z: Technical Summary for Customer Planning. However. Processing can measure performance.

the input traffic will be distributed across the different paths according to the bandwidth available on each of the paths. General rules for traffic isolation zones are: An N_Port can be a member of only a single TI zone. TI zone can be used for failover traffic from other TI zones when the ISLs fail. For FICON. An E_Port can be a member of only a single TI zone. and from ports that are not in a TI zone. the routing path is based on the Source ID (SID). this checking is done only against the current ADs’ zone database. Every time there is a change in the network. It is used to: Separate disk and tape traffic Select traffic for diverse ISL routes Guarantee bandwidth for mission-critical data Traffic isolation can operate with failover enabled or disabled. Traffic isolation The term “preferred path” for FICON traffic was replaced by the use of static routes and Traffic Isolation (TI) zones (not defined by the Fibre Channel standard). but can be utilized for Intermix Fabrics for FCP in a virtually-separated fabric. This feature is not supported for FICON Fabric. and Fibre Channel originator exchange ID (OXID). ensures reliability and availability even if a link in the trunk fails. Every exchange can take a different path through the fabric. maintains in-order delivery to ensure data reliability. Destination ID (DID). the input is redistributed across the available paths. Any traffic on ports that are not included in a TI zone follow the fabric shortest path first (FSPF) algorithm.level. only the source-destination policy is supported with guaranteed in-order-delivery of frames. and simplifies management by reducing the number of ISLs. Exchange-based routing requires the use of the Dynamic Load Sharing (DLS) feature. Dynamic Path Selection With Dynamic Path Selection (DPS). The choice of port-routing path is based only on the incoming port and destination domain. The balancing is based either on a source-destination or source-destination-exchange ID policy. Port-channel The Port-channel feature is an equivalent for link trunking on Cisco FICON Directors. Traffic Isolation allows data paths to be specified. DPS technology is currently not supported for FICON. If administrative domains (ADs) are configured. Lossless DLS Lossless Dynamic Load Sharing (DLS) allows you to rebalance trunk port paths without causing input/output (I/O) failures in cases where the end devices require in-order-delivery (IOD) of frames. 90 FICON Planning and Implementation Guide . When there are multiple paths to a destination. Port-based DLS DLS monitors link or trunk utilization to ensure load balancing. Frame-based trunking is the most effective way to utilize ISLs for FICON traffic while maintaining in-order-delivery (IOD) of frames inside the fabric. It is recommended that you either put all ports in TI zones or do not use TI zoning at all.

In most FICON environments.9.If multiple E_Ports are configured that are on the lowest-cost route to a domain. Port fencing allows the user to limit the number of errors that a port can receive by forcing a port offline when certain error thresholds are met. 4. The recommended thresholds are: Five errors per minute for CRC errors 25 errors per minute for Invalid Transmission Words Two errors per minute for Loss of Sync These settings are high enough to ignore occasional errors and transient errors due to recabling. tracking the utilization levels of disk subsystems. Port fencing Bad optics can cause errors to occur at a high enough frequency that error processing and sending and processing RSCNs can cause fabric performance problems. and resource recovery: QoS high. 4. Virtual channels create multiple logical data paths across a single physical link or connection.9 Evaluation tools To maintain business-critical mainframe infrastructures you need the right tools. dynamic profiling. Planning the FICON environment 91 . and expertise. load balancing is applied. and set an SNMP trap.9. traffic management. log an alert. experience. flow control mechanisms such as buffer credits). medium. you need additional performance data. Adaptive Networking can assign a Quality of Service (QoS) priority to individual data flows in each virtual fabric. port fencing is only set for CRC errors. it is difficult to conduct capacity planning or predict the impact of FICON migration and other changes. It anticipates congestion and dynamically makes adjustments in the fabric so that application traffic continues to flow. channels. only fencing the port and logging the alert are desired.8 Traffic management Virtual channel technology is the fundamental building block used to construct Adaptive Networking services. By default. host adapters. the alarms are set to fence the port. and RAID ranks. They allocate their own resources (queues. Without sufficient information. and Loss of Sync. Resource recovery prevents resource exhaustion and application disruption. If no other paths exist. For identifying performance bottlenecks. send and e-mail. Automatic rate limiting and advanced queuing priority algorithms are implemented to remove congestion and dedicate bandwidth or assign data flow to specific path. the dedicated E_Ports are used for other traffic also. For FICON environments. Adaptive Networking introduces four new networking services: Quality of Service. and low priority are assigned to applications using zoning (Source ID/Destination ID pairs on WWN base). Invalid Transmission Words. but low enough to stop problematic optics from causing fabric issues. Only source ports included in the zone are routed to zone E_Ports as long as other paths exist. Dynamic profiling provides knowledge about end-to-end resource use. Chapter 4.

IBM FICON aggregation study When planning to perform a capacity planning study for FICON channels. if additional channels are needed. long-term overview reporting with post-processor reports. refer to z/OS RMF Report Analysis. because no single vendor can provide the entire infrastructure. For System z. This provides online. Some reports to assist in your analysis are: Channel path activity report This report identifies each channel path by identifier and channel path type and reports on the channel utilization by server and individual LPAR. Device activity report This report provides activity and performance information for selected devices. SC33-7991. most SAN environments have more than one server vendor and multiple storage vendors. Disk controller evaluation . managing.IBM zCP3000 study – Channel evaluation . For further information. implementing. FICON Director activity report This report provides information about Director latency caused by port contention. you can provide the data and insight that are needed to effectively manage mainframe storage performance. I/O queuing activity report This report provides information about the I/O configuration and activity rate. zCP3000. The tools are listed and explained here: Processor evaluation . as well as optimize disk subsystem investments and fine-tune z/OS tape workloads. You can model.IBM Disk Magic™ study This is limited to the storage FICON channel port and port busy utilization numbers for workloads and configurations. use the IBM System z Capacity Planning tool. queue lengths.10 Prerequisites and interoperability Device interoperability within a SAN is very important. zCP3000 is designed to support FICON channel aggregation (select channel candidates for aggregation onto FICON channels).Using the suite of evaluation tools. and percentages when one or more I/O components were busy. IBM zHPF evaluation study This study is designed to quantify the amount of I/Os in an environment that appears to be zHPF-eligible. SAN vendors provide comprehensive interoperability matrixes and have extensive experience designing. and troubleshooting heterogeneous SANs. With this information. the z/OS Resource Measurement Facility (RMF) can be used to analyze your environment. the impact/benefit of zHPF can be assessed for a specific situation or workload. 92 FICON Planning and Implementation Guide . Shared device activity report This report gives you an overall performance picture of disk and tape devices that are shared between MVS systems in a sysplex. interactive performance monitoring or alternatively. 4. In fact.

refer to Chapter 2. For this reason. For FICON Director details. moves. “System z FICON technical description” on page 13.Directors providing connectivity for storage and servers must also have common services and settings to operate with stability and full performance. RSCNs. and changes Failing fabric merges due to slight differences in standard implementations Name server synchronization issues Security features implementation Special vendor features to enhance fabric performance E_Port connectivity in these situations is used to form a single fabric. refer to Chapter 3. FCR is not supported by FICON. “FICON Director technical description” on page 33. coordinating fabric information between heterogeneous platforms is still an issue for fabric interoperability. When interconnecting Directors. unmerged fabrics and to be managed independently.ibm. zoning. see: http://resourcelink. and performance measurement Slow fabric recovery during adds. IBM System Storage Web site: http://www-03. zoning database. FSPF.11 Physical connectivity The physical layer for FICON infrastructure is based on fiber optic cabling. which can lead to fabric connectivity issues. Planning the FICON environment 93 .ibm. Using Fibre Channel Routing is an option to allow secure communication between hosts and storage in two or more separate. node port states.html – IBM System Storage SAN768B Fiber Backbone Interoperability Matrix – IBM System Storage SAN384B Fiber Backbone Interoperability Matrix – Cisco MDS 9506. When planning this infrastructure you need to consider the following items: Existing physical layer of the data center Existing FICON components Planned FICON components Non-FICON components that also require fiber cabling Capacity planning of the data center’s physical layer Chapter 4. the name server.com/systems/storage/san/index. 9513 for IBM System Storage Directors Interoperability Matrix For more details regarding IBM System z qualification letters. and time-out value information must be shared. any cascaded FICON Directors must be from the same vendor. Native connectivity is required in the following situations: Migration between two platforms Port optimization within a fabric Temporary platform merge for infrastructure change Double vendor policies Issues introduced by the native connectivity include: Management interface inconsistencies in fabric topology. 9509. Even if vendors follow Fibre Channel standards.com 4. Vendor implementation of the same standard could be at different stages and include varying enhancements. For System z details.

which guarantees a high quality cable. In IBM terminology it is the Central Patching Location. multiple networks typically share a structured cabling infrastructure in a hierarchical star topology. local area networks (LANs). storage. We recommend using device-specific components. and it contains a large quantity of patch panels. remember to plan for a sufficient number of patch cords and length. the more stringent the requirements for the fiber and the connectors. placed in cabinets. When choosing the fiber media. we recommend using only qualified vendor fiber optic cabling components. scalability. Therefore. and SANs. The MDA is the central distribution point in the data center for telecommunication networks. We recommend using the OS2 fiber type for System z FICON LX features. moving. as well as multimode OM1. Also.The supported standard media for System z FICON features are single mode OS1 and OS2. With the increasing data rate on the link you need to utilize the latest fiber media types when designing the physical layer. and OM3. but their use could introduce a risk for interoperability and may not be supported by the equipment vendor. and high number of cycles for reuse. local distribution points. and Director environments of the data center. we recommend using the OM3 fiber type. and security in the data center infrastructure. The System z FICON SX features support multimode fiber. That means low attenuation. and changing devices inside a data center. All passive cross-connects. At the hub of the hierarchical star topology is the Main Distribution Area (MDA). They are delivered with a measurement protocol directly from the manufacturer. This provides a better starting point for cable routing inside the device rack. This type of approach also provides a highly scalable and flexible cabling infrastructure. Note: OM1 and OM2 fiber types are not recommended for laser-based systems. For the best performance and investment protection for multimode. It is the central cross-connect of the data center for all fiber optic and copper cables. according to the vertical and horizontal cable management 94 FICON Planning and Implementation Guide . for example MTP brackets designed for System z. This means the passive cross-connect points and traces are shared. This is important because the higher the bit rate on the channel. because of the lower attenuation compared to the OS1 fiber type. The highest flexibility and investment protection is with the single mode fiber. OM2. high return loss. The use of high-density cabling between the ZDA and the MDA reduces congestion and improves manageability. Data center cabling infrastructure To allow for adding. The MDA uses backbone or horizontal cabling to extend connectivity to the Zone Distribution Area (ZDAs) in the server. A unified media platform is also very important inside the data center for simplified operations and troubleshooting. modularity. it is important to view the FICON interconnections as an integral part of the overall fiber cabling infrastructure. When planning the routing of cables inside the cabinets. Qualified cabling components (connectors and cables) follow a standard color code. New fiber media types provide more flexibility and better performance. There are higher bandwidth fiber cables on the market today. avoid mixing cabling components from different vendors even if they are all qualified. and equipment outlets are based on high density patch panels with a unified connector type for better scalability.

or 4 Gbps may not be affected. MU. For more detailed information about fiber optic link testing and measurement. 2 Gbps. Refer to IBM Fiber Optic Cleaning Procedure. SY27-2604 for the procedure and materials required. The cable plant must satisfy the industry standard specification to minimize connector loss and reflections. The standard-based high density connectors are LC. We recommend using a unified connector type for the entire infrastructure.inside the cabinets. If you are experiencing excessive bit errors on a link (regardless of the data link rate).org http://www. Failure to do so can cause multiple outages during the operations of the devices. You have to consider the naming conventions for each environment specifically. oil. ease of use. Chapter 4. The cleaning is best performed by skilled personnel. pay extra attention to the power level alignment between the receiver and the transmitter on the channel between two components. keep in mind the format used in the FICON environment. This is the default for the IBM Fiber Transport System.tiaonline. We recommend using the standard-based ANSI/TIA-568-B. refer to the following URLs: http://www. Planning the FICON environment 95 . Low power levels do not necessarily mean a bad cable. and MTP. When creating a naming convention for System z. we provide a walkthrough of our FICON environment with a cascaded FICON Director topology in Appendix A. For additional information. The cleaning procedure may need to be performed more than once to ensure that all dust. when moving or laying cables. The FICON Express8 features utilize the existing single-mode and multimode cable plants. Coupling Links. The SCDC was especially developed for the mainframe fiber optic cabling environment. Dust. starting with ESCON. which in turn will impact performance. We strongly recommend you thoroughly analyze the fiber optic link to ensure that the cable plant specifications (total cable plant optical loss as well as connectors and splices return loss) are being met for that link length. the 8 Gbps channel is more sensitive to the condition of the cable plant. we recommend that you first clean and reassemble the connections. dirt. Each cabling infrastructure needs to be carefully documented. “Example: planning workflow” on page 243. The IBM high-density connector type is referred to as the small form factor duplex fiber optic connector for the data center (SCDC). and Open System Adapters).ibm. or oil is removed.com/services/siteandfacilities For an example of how you can use this decision-making process when planning a FICON environment. The most common source of cable plant optical link problems is associated with the various connections in the optical link. although lower link data rates such as 1 Gbps. refer to System z Maintenance Information for Fiber Optic Links (ESCON. When designing a fiber infrastructure.1-7 Method C. but rather the potential of port errors and resending of frames. FICON. and fiber link parameters. or defective connections may cause a problem with high speed channels such as 8 Gbps. When planning the optical interfaces. However. also consider the polarity management. number of channels. dirt. SY27-2597. There are various types of high density connectors available that offer differing parameters such as robustness.eu http://www. for ease of use and understanding by the operators.cenelec.

96 FICON Planning and Implementation Guide .

2009.Part 3 Part 3 Configuring the FICON environment This part provides you with a basic guide for configuring a FICON environment. All rights reserved. We give step-by-step guidance with descriptions of the software components and definitions that are needed to make the FICON environment operational. 97 . © Copyright IBM Corp. 2006. 2005.

98 FICON Planning and Implementation Guide .

and device definitions in IOCP Activating the IOCDS Configuring the storage control unit Varying the channel path and device online Verifying the configuration © Copyright IBM Corp. All rights reserved. Configuring a point-to-point topology This chapter describes all of the tasks that are required to configure and define a FICON environment on a point-to-point topology. 2009.5 Chapter 5. control unit. Establishing a point-to-point topology involves the following items: Defining FICON channel. 99 . 2005. 2006.

Our discussion is based on an infrastructure that is already built and installed. In our case. including configuring the attached storage control unit. The two host adapters installed in the DS8000 are longwave (LX) laser. An IBM System Storage DS8000 will be connected to the z10 server via CHPID 7F and CHPID 83. which corresponds to the LPAR name A29. The system name running in this partition is SC33. including the physical installation of the following items: System z server DS8000 storage system Fiber cabling infrastructure 5. As a standard feature of the z10 servers. The configuration is shown Figure 5-1. zHPF protocol will be used for data transfer to the DS8000 storage device. Two logical control 100 FICON Planning and Implementation Guide .1 Description of our environment Based on the considerations explained in Chapter 4. using PCHID 5A3 (CHPID 7F) and PCHID 5D3 (CHPID 83).5. Port numbers 0132 and 0331 in the DS8000 are used to connect to the server.1 Establishing a point-to-point topology This section describes the tasks that are required to configure a FICON point-to-point topology on a System z server. Two FICON channels are defined as shared and spanned across CSS2 and CSS0 (only CCS2 is shown). The channels are FICON Express8 LX features (FC 3325). all the required infrastructure is already installed in our environment. we are going to build a FICON point-to-point configuration that consists of a z10 EC server and a DS8000 storage device. z10 EC SCZP201 LPAR SC33 A29 z/OS V1R10 FICON Express8 LX CSS2 PCHID 5A3 CHPID 7F DS8000 CU# D000 LX 0132 D0xx FICON Express8 LX PCHID 5D3 CHPID 83 LX 0331 D1xx CU# D100 * All cable connectors are LC Duplex type Figure 5-1 Our point-to-point configuration The z10 EC server (SCZP201) has one LPAR defined and activated.1. The operating system running in this partition is z/OS V1R10. “Planning the FICON environment” on page 59.

and devices on a System z server Information about defining the channel paths.units (D000 and D100) are defined. Both hardware and software requirements must be checked. CU. which have devices D000-D0FF and D100-D1FF assigned. CU. The DS8000 will have the zHPF feature enabled. see “System z FICON feature support” on page 25. Figure 5-2 Main steps for configuring and verifying a FICON point-to-point configuration Verification checklist Before configuring the point-to-point topology shown in Figure 5-1 on page 100. Tasks Figure 5-2 shows the main steps required to define and activate a FICON point-to-point configuration. All steps in the checklist must be finished and corrected (if required) to ensure a smooth and successful configuration of the topology. Verify configuration Information about how to verify that your actual configuration matches the desired configuration is given in “Verifying the installation” on page 110. The fiber optic cables have an LC Duplex connector at both ends to connect to the z10 FICON Express8 channels and to the DS8000 host adapters. Follow the verification checklist to ensure that all HW and SW prerequisites are met.2 miles) is supported by the longwave laser (LX) feature when using 9µm single mode (SM) fiber optic cable. and storage device in IOCDS” on page 103. Check that the appropriate FICON features are available on the System z server. Verification checklist Define channel. Go to “Verification checklist” on page 101.2 Tasks and checklist This section provides an overview of required tasks and the checklist that we used to make sure all hardware and software prerequisites were met before starting the configuration tasks. 5. Plug fiber cables Information about fiber optic cables and plugging rules to achieve the desired configuration is given in “Connecting the fiber optic cables” on page 108. Configure storage CU The configuration tasks for DS8000 storage system are described in “Configuring the IBM Storage System DS8000” on page 108. control units. the following list was checked.1. and devices is given in “Defining the channel. A maximum unrepeated distance of 10km (6. For details about each feature. Configuring a point-to-point topology 101 . FICON Express2 LX FC 3319 Chapter 5.

7 with the IBM Lifecycle Extension for z/OS V1. Note: FC 0709 and FC 7092 must be ordered to obtain the license key to enable zHPF support in a DS8000 storage controller. It explains what needs to be done to get the FICON channels.5µm multi-mode (MM) fiber optic cable is required to support shortwave laser (SX).7 (5637-A01) is required to support the FICON Express8 feature in the z10 server. and the storage device online and operating. A 9µm single mode (SM) fiber optic cable is required to support longwave laser (LX) for a maximum distance of 10km (6.1.16. Check that the correct fiber optic cables are available to connect the DS8000 storage controller to the z10 server. Check that the zHPF feature is enabled in the DS8000. Check that FC 0709 and FC 7092 are installed to support zHPF. Check the 2097DEVICE PSP bucket for the latest information about FICON Express8 support on the operating system. The steps include: Defining the channel. 5. See “System z FICON feature support” on page 25 for the maximum supported distance depending on cable type and speed. z/OS V1.1. FC 3324 SX FC 3318. Check DS8000 storage hardware requirements to support FICON longwave (LX) or shortwave (SX) connectivity. check the System z operating system requirements. CU.SX FC 3320 FICON Express4 LX FC 3321. A 50µm or 62. For example. FC 3323. FC 3322 FICON Express8 LX FC 3325 SX FC 3326 If using the FICON Express8 feature.0 or higher to support the zHPF feature. Note: All fiber optic cables used in a link must be of the same type. they must be either all single mode or all multi-mode fiber optic cables.3 Getting started This section describes all the tasks that are required to reach the designated configuration shown in Figure 5-1 on page 100. and storage device in IOCDS Activating the changes (IOCDS) and configuring online Configuring the IBM Storage System DS8000 Connecting the fiber optic cables Configuring the CHPIDs online Verifying the installation 102 FICON Planning and Implementation Guide .2 miles). storage control units. Check that DS8000 firmware is at level 6. An LC duplex connector is required at both ends of the fiber optic cables to connect to the System z server FICON adapter and the DS8000 host adapter.

If changes to the I/O definitions are performed dynamically. The TYPE of the channel is specified as FC (native FICON). Defining the channel.PCHID=5A3. “System z” on page 78.PCHID=5D3. Because no system outage is required for dynamic changes. Example 5-1 CHPID definition for point-to-point configuration CHPID PATH=(CSS(2). Chapter 5. the definition data is stored in the I/O definition file (IODF) data set. we had an active partition in a System z server running z/OS with HCD. The same definition rules apply to the definition of CHPID 83.8. Configuring a point-to-point topology 103 . LPAR names specified in ICODS correspond to system name SC33. Changes or additions of channel paths. and storage device in IOCDS There are several tools available that you can use to define and configure a FICON environment on a System z server. HCD was used to define the CHPIDs. Any time you make changes to the I/O configuration. The following examples show the CHPID. those changes are directly updated into the Hardware System Area (HSA) of the server. “Configuration and definition tools” on page 255.PARTITION=((A29).(=)). The PCHID 5A3 has been merged to the IOCDS by the CMT.SHARED. The newly created IOCDS will be sent to the Support Element of the system server. First we explain how the CHPIDs are defined. Only statements and keywords relevant to our desired configuration are explained. CUs. See Example 5-1 for the CHPID statement and its keywords.* TYPE=FC The PATH keyword in the CHPID statement defines CHPID 7E in CSS 2 as SHARED. The Support Element in the System z server stores the IOCDS on its built-in HDD.83).7F).PARTITION=((A29). SC33-7988. The definition tools and an explanation of the statements and keywords used in the IOCP are described in Appendix C. refer to 4. Refer to HCD User’s Guide.SHARED. and devices for the sample configuration we describe in this section. this is the preferred method for changing I/O definitions. We used HCD to create. for details about all available IOCP statements and keywords. Refer to IOCP User’s Guide.For our scenario. CU. and IODEVICE statements when all the definition tasks based on the configuration shown in Figure 5-1 on page 100 are finished.(=)). The IOCDS stored in the Support Element is loaded into the server’s storage during a Power-on Reset (POR) of the System z server. The HSA is a separate storage area in the System z server which holds all I/O specifications and provides status information for all defined I/Os. CUs.* TYPE=FC CHPID PATH=(CSS(2). for detailed information regarding the use of HCD. or devices can be performed dynamically. save. SB10-7037.1. CNTLUNIT. and activate our I/O definitions. The LPAR name (A29) is specified in the PARTITION keyword to use CSS 2 to access CHPID 7F. The operating system uses the HSA to manage the I/O subsystem. After all definition tasks are completed. For more details about System z channel subsystem characteristics. the IOCDS has to be updated to ensure consistency between the software and the microcode in the hardware. which assigned CHPID 7F to PCHID 5A3.

With the CHPIDs defined. Defining a FICON Channel-to-Channel connection In addition to storage control units and storage devices.(CSS(1). Example 5-2 CNTLUNIT definition for point-to-point configuration CNTLUNIT CUNUMBR=D000. Example 5-3 displays the IODEVICE statement and keywords.**). Two asterisks (**) are specified in the LINK keyword for point-to-point connectivity.(CSS(1). The same definitions are done for devices associated to control unit D100. we have defined FICON Channel-to-Channel (FCTC) CUs and devices to allow data exchange between LPARs. * UNITADD=((00. In addition.UNIT=3390B ADDRESS=(D171.7E.5C.5D).0E. * CUADD=0.0E.5D).(CSS(2).54. * LINK=((CSS(0). the next step is to define the devices owned by the CUs.54.28.STADET=Y.55. For control unit D100.UNIT=3390A The IODEVICE statement specifies the characteristics of devices D000 and D100.UNIT=3390A ADDRESS=(D100. Example 5-3 IODEVICE definition for point-to-point configuration IODEVICE IODEVICE IODEVICE IODEVICE ADDRESS=(D000.0A.58.51.28. * CUADD=1.5C). starting at device address D000.0A)).UNIT=2107 CNTLUNIT CUNUMBR=D100.7F.143).* 51.58.55.55.28.STADET=Y.50.28.CUNUMBR=(D100). * UNITADD=((00.5C)).UNIT=2107 There are two control units defined by the CUNUMBR keyword: D000 and D100.50.0E. For control unit D000 there are 113 devices defined.51.66CF. * PATH=((CSS(0).0A.83).58.113).0E.55.(CSS(1).7E. For more information about considerations regarding FCTCs.82.58.5C).59.2C. refer to Appendix G.54.5D)).(CSS(3).28.2C.256)).**).0E)).2C.51.CUNUMBR=(D100).0A.54.50.* 50. we next show how the CUs attached to the CHPIDs are defined.CUNUMBR=(D000).STADET=Y.(CSS(2).6688.**.**.2C.59.28.0A). 143 Hyper PAV alias devices are defined starting at device address D071 and D171.66CF. No link address is required for CHPID 7F and 83.CUNUMBR=(D000). “Adding FICON CTC connections” on page 305. * LINK=((CSS(0).* 2C. The logical CU image number 0 is specified for control unit D000.113). The LINK keyword is required because some other FICON channels have access to the CUs through a FICON Director.6688.2C.(CSS(3).(CSS(2).STADET=Y. CHIPDs 7F and 83 have access to CUs D000 and D1000 because both CHPIDs (among others) are specified in CSS2 in the PATH keyword for both CUs. Example 5-2 displays the CNTLUNIT statement and keywords.28.(CSS(1).2C. 104 FICON Planning and Implementation Guide .0A.(CSS(2).5D.0E). * PATH=((CSS(0). the logical CU image number 1 is specified in the CUADD keyword.0E. After the CHPIDs and the CUs are defined.83).UNIT=3390B ADDRESS=(D071.7F.2C.* 28. These are base devices in the storage unit.(CSS(3).0A.(CSS(3).82.143).0A).256)).59.0E).59.

A23)).UNIT=FCTC IODEVICE ADDRESS=(5034.A24)). Configuring a point-to-point topology 105 . CUADD=24. Example 5-4 shows the definition of FCTC control units and devices.7E)).004)). There is a direct connection between CHPID 7E and CHPID 82 by a fiber optic cable. LPAR A24 will communicate via a FICON channel to CU 4044. To simplify the illustration only four CUs and two LPARs are shown in this sample configuration.PATH=((CSS(2).UNIT=FCTC IODEVICE ADDRESS=(4044. STADET=Y.CUNUMBR=(4044).PARTITION=((CSS(2). which is assigned to LPAR A23.UNITADD=04.UNITADD=04. z10 EC SCZP201 LPAR SC30 A23 z/OS V1R10 FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E CU# CU# CU# CU# 4044 5044 4034 5034 LPAR SC31 A24 z/OS V1R10 FICON Express8 LX CSS2 PCHID 5D2 CHPID 82 * All cable connectors are LC Duplex type Figure 5-3 FCTC configuration point-to-point Both LPARs have access to CHPID 7E and 82 because the CHPIDs are defined as shared. STADET=Y.004).UNITADD=((04. Each LPAR will have access to a logical control unit and their logical devices defined in the corresponding LPAR.UNIT=FCTC * * * * One control unit is defined for each LPAR: CU 5034 for LPAR A23.Based on those considerations.82)). which is assigned to LPAR A24. and CU 4044 for LPAR A24.CUNUMBR=(5034).PARTITION=((CSS(2).004). we are going to configure FCTCs in a point-to-point FICON environment. Figure 5-3 illustrates the desired FCTC configuration. Example 5-4 FCTC definitions point-to-point CNTLUNIT CUNUMBR=5034. There are four devices specified for each LPAR (LPAR A23 and LPAR A24).UNIT=FCTC CNTLUNIT CUNUMBR=4044. The PARTITION keyword in the IODEVICE statement specifies the logical partition that can access the device using the FICON channel.PATH=((CSS(2). LPAR A23 will communicate via a FICON channel to CU 5034.004)).UNITADD=((04. This allows data traffic to flow between both LPARs using the corresponding devices: SC30 SC31 4044 <---> 5034 Chapter 5. CUADD=23. although more than four CUs could be defined.

4045 <---> 5035 4046 <---> 5036 4047 <---> 5037 The rules we followed for device numbering are described in “FCTC device numbering scheme” on page 306. SC33-7988. Figure 5-5 on page 107 shows the Activate or Process Configuration Data menu. To simplify the illustration. Refer to HCD User’s Guide. Activating the changes (IOCDS) and configuring online After the definition of the desired configuration is finished. only two logical CUs are shown. for detailed descriptions of all activation procedures. 106 FICON Planning and Implementation Guide . Figure 5-4 illustrates the logical view of the FCTC configuration and the data path between LPARs. the data must be saved into the IODF data set and the IOCDS file on the Support Element. LPAR A23 sends data to A24 via CU 5034. Up to four IOCDS files can be stored in a System z server. On the HCD start menu. The reverse applies to LPAR A24 when data is sent to or received from A23. This allows to you plan and define future I/O configurations and store them on the Support Element of the server. LPAR A23 receives data from A24 via CU 4044. LPAR A23 Data transfer CU 4044 7E 82 CU 5034 Data transfer LPAR A24 Figure 5-4 FCTC data transfer (point-to-point) Data is transferred over the FICON link in both directions between the CUs and the logical partitions (LPARs). select option 2 (Activate or process configuration data). The tasks that are required to save definition data and activate changes dynamically are performed via HCD.

Figure 5-5 HCD Activate or Process Configuration Data menu The options used to activate the changes are: Option 1. enter the D IOS. Changes can be done at any time and saved in an IODF data set. Keep in mind. The IOCDS is built using the definitions saved in the IODFxx data set. Configuring a point-to-point topology 107 . To verify which I/O definitions (IODFs) are currently active. that a POR causes a system outage. The newly-built IOCDS based on the I/O definition in the IODF should be sent to the CPCs Support Element and marked for use at the next POR. The response to the command is shown in Example 5-5. however. Another way to activate changes in I/O definitions is to perform a POR at the CPC with the newly-built IOCDS. This ensures that the desired IODF used by the operating system image matches the IOCDS that is loaded into the HSA. The suffix of the IODF data set is used in the IPL parameter when the z/OS image is brought up. Chapter 5. then we suggest that you plan ahead to make any changes in I/O definitions so that the new IOCDS is ready for use at the next POR. where xx in the data set name is a suffix used to identify the IODF among other IODFs which may already exist. Activate or verify configuration dynamically The Build production I/O definition file (IODF) function (option 1) saves definition data into an IODFxx dataset. Build ICODS Option 6. Using Activate or verify configuration dynamically (option 6) allows you to dynamically activate changes in the I/O definitions. if a POR is required. This is the preferred method for activating changes. There is no Power-on Reset (POR) required because changes are directly updated to the HSA in the CPC. For this reason. This ensures that the newly-built IOCDS will be used the next time the CPC is activated. Using Build IOCDS (option 2) allows you to build an IOCDS and save it to the CPC.CONFIG command at the z/OS operator console. Build production I/O definition file Option 2.

IODF01 CONFIGURATION ID = TEST2094 EDT ID = 01 TOKEN: PROCESSOR DATE TIME DESCRIPTION SOURCE: SCZP201 09-05-11 20:50:41 SC33 IODF01 ACTIVE CSS: 2 SUBCHANNEL SETS CONFIGURED: 0.CONFIG command IOS506I 15.CONFIG command with the data of the ID statement in the IOCDS. All fiber optic cables must have LC duplex connectors at both ends.20.1). as listed here: PCHID 5A3 to DS8000 host adapter port 0003 PCHID 5D3 to DS8000 host adapter port 0242 Make sure the fiber optic cable ends are cleaned before plugging them into the transceivers. The configuration steps are based on the configuration shown in Figure 5-1 on page 100.LSYSTEM=SCZP201.Example 5-5 D IOS.'20:50:41'.IODF01 .'IODF01') The suffix of the IODF data set shown in the IOCDS ID statement and displayed by the D IOS command should be the same number (that is. to the FICON channel and the CU ports. 1 CHANNEL MEASUREMENT BLOCK FACILITY IS ACTIVE Example 5-5 shows that the IODF01 data set is currently active in z/OS image SC33. Compare the information provided by the D IOS. All fiber links between the FICON channel and the CU host adapters use the longwave (LX) type of laser. CSS2 is the active CSS in system SC33.'SC33'.2009-05-11 20:50'. when recabling or for problem determination).00800006991E2094205041680109131F00000000. Also ensure that all fiber optic cables are labeled and documented for later use (for example. shown in Figure 5-1 on page 100. shown in Example 5-6. Plug the fiber optic cable connectors to the designated ports. Connecting the fiber optic cables After all definition and configuration tasks are completed. “Configuring the DS8000 for FICON” on page 259 to configure the host adapter ports. For planning and implementation information. Follow the procedure described in Appendix D. This will prove that the IOCDS was built from the current active IODF data set and that the latest changes and new IO definitions were successfully activated. * TOK=('SCZP201'. fiber optic cables have to be plugged into the FICON channels and CU host adapters. * SYSTEM=(2097. Example 5-6 ID statement in IOCDS ID MSG1='IODF01'.MSG2='SC33. 108 FICON Planning and Implementation Guide . Configuring the IBM Storage System DS8000 The host adapter ports in the DS8000 storage control unit have to be configured for native FICON protocol. SG24-6786. using a 9µm single mode (SM) fiber optic cable is recommended to achieve maximum performance and distance.'09-05-11'. For the LX type of laser. It also provides information about the CSSs used in this z/OS image. refer to IBM System Storage DS8000: Architecture and Implementation. IODF01).* 00000000.05 I/O CONFIG DATA 345 ACTIVE IODF DATA SET = SC33.

DEVICE STATUS FOR CHANNEL PATH 7F 0 1 2 3 4 5 6 7 8 9 A B C D 0D00 + + + + + + + + + + + +@ +@ + 0D01 + + + + + + + + + + + + + + 0D02 + + + + + + + + + + + + + + . It provides information about the channel and the attached devices. However.922. Example 5-9 D M=CHP(7F) D M=CHP(7F) IEE174I 11.58..9. Note that even if all the definitions in the IOCDS are correct. Example 5-8 Enabling zHPF D IOS. 0DFD HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFE HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFF HA HA HA HA HA HA HA HA HA HA HA HA HA HA SWITCH DEVICE NUMBER = NONE ATTACHED ND = 002107. enter SETIOS ZHPF. the zHPF facility will be reset to the default (disabled).ZHPF at the z/OS console to display zHPF settings for the z/OS image. Configuring the CHPIDs online After you successfully activate the new I/O definitions.YES. Example 5-7 CF CHP(7F) CF CHP(7F) IEE502I CHP(7F).11.59 DISPLAY M 290 CHPID 7F: TYPE=1A. DESC=FICON POINT TO POINT.75. “High Performance FICON” on page 88 for considerations regarding how to exploit zHPF on System z servers. Configuring a point-to-point topology 109 . If the CHPID is offline. after the next system IPL. Enter D IOS. The command output is shown in Example 5-9. enter D M=CHP(7F) on the system console to display the current status of CHPID 7F.For further information and considerations regarding fiber cabling and documentation. Now you can query the status and functional details of the channel by entering D M=CHP(7F) at the operator console. as shown in Example 5-8. The channel status should change to ONLINE. as shown in Example 5-7.27. refer to 4. the channel status might be offline. “Documentation” on page 62. change the FCX parameter in the SYS1.ONLINE IEE712I CONFIG PROCESSING COMPLETE To achieve the best performance on the FICON channel. you can configure the channel path online and try to communicate to the devices.23 ZHPF FACILITY 021 HIGH PERFORMANCE FICON FACILITY IS ENABLED Using the SETIOS ZHPF=YES command enables zHPF temporarily.0000000BALB1 ONLINE E + + + F + + + HA HA HA HA HA HA Chapter 5. Refer to 4. This enables zHPF.PARMLIB member IECIOSxx to FCX=YES.IBM. If zHPF is disabled.ON to configure the channel online. “Physical connectivity” on page 93 and 4..6.2. enter CF CHP(7F). make sure that zHPF is enabled. To permanently enable zHPF for z/OS. therefore.ZHPF RESPONSE=SC33 IOS630I 13.

0000000BALB1. ZHPF 83 . Information is also displayed about the attached devices and the facilities (for example. Display the list of channels installed on the CPC. the control unit number.IBM. where xxxx is any device number. and the functions supported by the device (for example. Search the list for the desired PCHID (5A3) and double-click the PCHID icon. SCZP201). To verify that communication to the attached devices is working properly.900.0000 SCP DEVICE NED = 002107.08. and to check the status of channels and devices. MIDAW and zHPF) is shown. First. This verification should be done to ensure that the fiber optic cables are correctly plugged to the designated ports in the System z server and the DS8000. Example 5-10 D M=DEV(D000) D M=DEV(D000) IEE174I 12.75.75. 110 FICON Planning and Implementation Guide . device type).0132 SCP TOKEN NED = 002107. check the status of PCHID 5A3 in the z10 server.0000000BALB1.0000 HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW. enter D M=DEV(xxxx). DEST LINK ADDRESS 2C 2C 0A 0A 6688 66CF 0D PATH ONLINE N N N N N N Y CHP PHYSICALLY ONLINE N N N N N N Y PATH OPERATIONAL N N N N N N Y MANAGED N N N N N N N CU NUMBER D000 D000 D000 D000 D000 D000 D000 MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107.PHYSICAL CHANNEL ID = 05A3 FACILITIES SUPPORTED = ZHPF Example 5-9 on page 109 shows that CHPID 7F is online and operating in a point-to-point configuration. you can verify that the current configuration matches the desired configuration shown in Figure 5-1 on page 100.900. Verifying the installation After all the configuration tasks in the z10 server and the DS8000 storage are complete. zHPF) supported by the channel.0000000BALB1. From any HMC where the desired CPC (server) is defined (for example. start a Single Object Operation (SOO) to that CPC.. This will display the PCHID details panel shown in Figure 5-6. Information about the device (for example..IBM.900. For example.IBM.75. to check the status of device D000.55 DISPLAY M 296 DEVICE D000 STATUS=ONLINE CHP 50 54 58 5C 7E 82 7F ENTRY LINK ADDRESS 0F 0F 1B 1B 6503 651B . 0D Y Y Y N D000 The response to D M=DEV(D000) displays all available paths to the devices and their status. enter D M=DEV(D000) on a z/OS console (see Example 5-10). which is a storage device.

The CHPID assigned to PCHID 5A3 is shared. This proves that PCHID 5A3 has CHPID 2. Figure 5-7 CHPID details The CHPID Details panel provides information that is similar to the PCHID details panel. Chapter 5. The PCHID type is FICON Express8. The owning image of PCHID 5A3 is A29. shown in Figure 5-7.7F is assigned to PCHID 5A3.7F assigned and that image A29 can access the channel. Repeat these checks on other channels that were recently configured. To display the CHPID details.Figure 5-6 PCHID details Essential information for PCHID 5A3 is shown on the PCHID Details panel: The PCHID status is Operating. Configuring a point-to-point topology 111 . you must correct the definitions in the IOCDS. select the channel list for an operating system image from the Support Element workplace.CHPID 2. CSS. Similar information can be found on the CHPID details panel. Important: If any of the data displayed on the PCHID or CHPID detail panel does not match the desired configuration. Notice that the information for PCHID 5A3 (CHPID 2.7F) provided on the details window matches the designated configuration.

as shown in Figure 5-9 on page 113.Next. which provides information about the node attached to the FICON channel. click the Channel Problem Determination button. where you can select which information you want to display (see Figure 5-8). Figure 5-8 Channel Problem Determination panel Select Analyze channel information and click OK. This will display the Channel Problem Determination panel. The Analyze Channel Information window is displayed. check that the FICON channels are connected to the correct host adapter port in the DS8000 storage controller. 112 FICON Planning and Implementation Guide . On either the PCHID details or CHPID details panel.

none of the displayed information is valid. that the tag value is provided by the attached device during link initialization and may have different meanings. The Tag provides information about the port number of the attached node. and may be used to prove that the channel is connected to the correct FICON adapter port if the WWN or WWP of the attached device is known. however. Check that the Type/model information as well as the serial number (Seq. Be aware. examine the Tag field for each node. Next. the four digits represent the port number (0132). The lower left side displays information about the node in the System z server. For the attached node. If any other status is shown. If the node status is not Valid or the Tag value and WWPN value are not correct. The lower right side displays information about the attached node. Important: Make sure that the Node status for both nodes is displayed as Valid. The two right-most digits of the Tag value represent the CHPID number for the channel node (7F). number) is as expected. Chapter 5. depending on the vendor. Configuring a point-to-point topology 113 . The World Wide Node Name (WWN) and the World Wide Port Name (WWP) are also shown for each port. check the fiber optic cable link between the z10 server and the FICON Director to ensure that it is plugged correctly.Figure 5-9 Analyze Channel Information Information about the nodes is displayed at the bottom part of the window.

If the displayed values are not as expected. Figure 5-10 Analyze Serial Link status The Analyze Serial Link Status window provides status information about the link to the control unit images defined in the IOCDS. select Analyze Serial Link Status and click OK. The Analyze Serial Link Status window is displayed. shown in Figure 5-8 on page 112. If the link status Initialization Complete is not shown. On the Channel Problem Determination panel. 114 FICON Planning and Implementation Guide .In our scenario we are now sure that PCHID 5A3 has connectivity to port number 0132 at the DS8000 host adapter. Figure 5-10 shows a link status of Initialization Complete for all defined CU images on CPID 7F. After completing the preceding steps and proving that the physical path to the DS8000 storage controller and the logical definitions of the link are correct. check that the path to the control unit image is initialized correctly and properly defined. a link address of 0D is displayed by default for a point-to-point FICON link. the fiber optic cables may not be plugged correctly and must be checked. and that the fiber optic cable link to the CU has the correct cable type and plugging. Scroll through the list of CU images and check that the status for all CUs displays as Initialization Complete. Although there is no FICON Director attached. you must check that the ports in the CU are correctly configured. as shown in Figure 5-10. which matches our desired configuration (see Figure 5-1 on page 100).

and device definitions in IOCP Activating the IOCDS Configuring the storage control unit Varying the channel path and device online Verifying the configuration © Copyright IBM Corp. 2009. All rights reserved. 2006. 115 . control unit. 2005. Establishing a switched point-to-point topology involves the following: Defining FICON channel. Configuring a switched point-to-point topology This chapter describes all the tasks that are required to configure and define a FICON environment on a switched point-to-point topology.6 Chapter 6.

Both LPARs have access to the two FICON channels. we are going to build a FICON switched point-to-point configuration. all the infrastructure is already installed in our environment. The operating system running in both partitions is z/OS V1R10. which consists of an IBM z10 EC server.1. an IBM System Storage SAN768B FICON Director.1 Description of our environment Based on considerations explained in Chapter 4. including the physical installation of the following: System z server DS8000 storage system FICON Director Fiber cabling infrastructure 6. and an IBM DS8000 System Storage device. Two FICON channels are defined as shared and in CSS2.1 Establishing a switched point-to-point topology This section describes the tasks that are required for configuring a FICON switched point-to-point topology on a System z server. zHPF protocol will be used for data transfer to the DS8000 storage device. which correspond to the LPAR names A23 and A24. including setup of a FICON Director and configuring the attached storage controller. respectively. “Planning the FICON environment” on page 59. In our case. As a standard feature of the z10 server. Our discussion is based on an infrastructure that is already built and installed. The configuration is shown Figure 6-1. 116 FICON Planning and Implementation Guide . The system names running in these partitions are SC30 and SC31. z10 EC SCZP201 LPAR SC30 (A23) z/OS V1R10 FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E System Storage SAN768B Port 03 Port 28 DS8000 CU# D000 LX 0003 D0xx Switch # 65 Switch @ 65 LPAR SC31 (A24) z/OS V1R10 FICON Express8 LX CSS2 PCHID 5D2 CHPID 82 Port 1B Port 3F LX 0242 D1xx CU# D100 * All cable connectors are LC Duplex type Figure 6-1 Switched point-to-point configuration The z10 EC server (SCZP201) has two LPARS defined and activated. The channels are FICON Express8 LX features (FC 3325) using PCHID 5A2 (CHPID 7E) and PCHID 5D2 (CHPID 82).6.

The fiber optic cables have an LC Duplex connecter at both ends to connect to the z10 FICON Express8 channels.2 miles) is supported by the longwave laser (LX) feature when using 9µm single mode (SM).The SAN768B FICON Director has connectivity to the z10 server on port 03 and port 1B. Define channel. CU. The switch number (Switch #) and the switch address (Switch @) are both set to 65. Information about defining the channel paths. The DS8000 will have the zHPF feature enabled. 6.1 “Configuration flowchart” on page 152. All required steps to set up the FICON Director are shown in 8. The two host adapters installed in the DS8000 are longwave (LX) lasers. Tasks Figure 6-2 shows the main steps required to define and activate a FICON switched point-to-point configuration. Go to “Verification checklist” on page 118. and to the DS8000 host adapter. Configure FICON director Plug fiber cables Information about fiber optic cables and plugging rules is given in “Connecting the fiber optic cables” on page 126. and devices is given in “Defining the channel. A maximum unrepeated distance of 10 km (6. Figure 6-2 Main steps for configuring and verifying a FICON switched point-to-point configuration Chapter 6. and device in IOCP” on page 119. Although the switch number is specified in the IOCP. Verify configuration Information about how to verify that your actual configuration matches the desired configuration is given in “Verifying the installation” on page 128. Verification checklist Follow the verification checklist to ensure that all hardware and software prerequisites are met. Port number 0003 and port number 0242 are used to connect to the FICON Director. to the ports in the FICON Director. CU. the switch address is the Domain ID specified in the FICON Director. and connectivity to the DS8000 storage system on port 28 and 3F. All the ports in the FICON Director are longwave (LX) ports.1. Configuring a switched point-to-point topology 117 . Two logical control units (D000 and D100) are defined and have devices D000-D0FF and D1000-D1FF assigned.1.2 Tasks and checklist This section provides an overview of the required tasks and a checklist that we used to make sure all hardware and software prerequisites were met before starting the configuration tasks. control units. and devices on a System z server Configure storage CU The configuration tasks for a DS8000 storage system are described in “Configuring the IBM Storage System DS8000” on page 125.

the following list was checked.7 (5637-A01) is required to support the FICON Express8 feature in z10 server. For details about each feature. FC 3323. Check that the appropriate FICON features are available on the System z server. FC 3324 SX FC 3318. Check the 2097DEVICE PSP bucket for the latest information about FICON Express8 support on the operating system. Check that FC 0709 and FC 7092 are installed to support zHPF. refer to “System z FICON feature support” on page 25. Note: FC 0709 and FC 7092 must be ordered to obtain the license key to enable zHPF support in a DS8000 storage controller. Check that the correct fiber optic cables are available to connect the FICON Director to the System z server. FC 3322 FICON Express8 LX FC 3325 SX FC 3326 If using FICON Express8 features. Check that the zHPF feature is enabled in the DS8000. the ports in the FICON Director.1. A 9µm single mode (SM) fiber optic cable is required to support longwave laser (LX) for a maximum distance of 10km (6. An LC duplex connector is required at both ends of the fiber optic cables to connect to the z10 server FICON channel.5µm multi mode (MM) fiber optic cable is required to support shortwave laser (SX). LX SFPs SX SFPs Check DS8000 storage hardware requirements to support FICON longwave (LX) connectivity to the z10 server.0 or higher to support the zHPF feature. See “System z FICON feature support” on page 25 for the maximum supported distance depending on cable type and speed. A 50µm or 62.7 with the IBM Lifecycle Extension for z/OS V1. z/OS V1. FICON Express2 LX FC 3319 SX FC 3320 FICON Express4 LX FC 3321. and the DS8000 storage to the FICON Director. Hardware and software requirements must be checked.16. Check that the DS8000 firmware is at level 6. 118 FICON Planning and Implementation Guide . Check that the number and types of FICON Director ports match the configuration requirements. and the DS8000 host adapter ports. check System z operating system requirements.2 miles).Verification checklist Before configuring the switched point-to-point topology shown in Figure 6-1 on page 116. All steps in the checklist must be finished and corrected (if required) to ensure a smooth and successful configuration of the topology.

The operating system uses the HSA to manage the I/O subsystem.3 Getting started This section describes all the tasks that are required to achieve the designated configuration shown in Figure 6-1 on page 116. CUs. Changes or additions of channel paths. and device in IOCP There are several tools available that you can use to define and configure a FICON environment on a System z server.1. CUs. storage control units. Refer to HCD User’s Guide. 6. for detailed information regarding the use of HCD. Chapter 6. The HSA is a separate storage area in the server which holds all I/O specifications and provides status information for all defined I/Os. The following examples show the CHPID. and devices for the sample configuration we describe in this section. this is the preferred method for changing I/O definitions. The newly created IOCDS will be sent to the Support Element of the System z server. Only statements and keywords relevant to our configuration are explained.8. and IODEVICE statements if all the definition tasks based on the configuration shown in Figure 6-1 on page 116 are finished. Configuring a switched point-to-point topology 119 .Note: All fiber optic cables used in a link must be of the same type. FICON Director. the IOCDS has to be updated to ensure consistency between the software and the microcode in the hardware. The IOCDS stored in the Support Element is loaded into server storage during a Power-on Reset (POR) of the System z server. we had an active partition in a System z server running z/OS with HCD. Because no system outage is required for dynamic changes. the changes are directly updated into the Hardware System Area (HSA) of the System z server. they must be either all single mode or all multi-mode fiber optic cables. and storage device in IOCP Activating the changes (IOCDS) and configuring online Configuring the IBM Storage System DS8000 Configuring the FICON Director Connecting the fiber optic cables Configuring the CHPIDs online Verifying the installation For our scenario. and activate our I/O definitions. The steps described are: Defining the channel. CNTLUNIT. Defining the channel. SC33-7988. We used HCD to create. After all definition tasks are completed. The definition tools and an explanation of the statements and keywords used in the IOCP are described in Appendix C.1 “System z” on page 78. save. or devices can be performed dynamically. For example. SB10-7037. If changes to the I/O definitions are performed dynamically. the definition data is stored in the I/O definition file (IODF) data set. Any time you make changes to the I/O configuration. The Support Element in the server stores the IOCDS on its built-in HDD. For more details about System z channel subsystem characteristics. refer to 4. HCD was used to define the CHPIDs. It explains what needs to be done to get the FICON channels. Refer to IOCP User’s Guide. “Configuration and definition tools” on page 255. CU. CU. and the storage device online and operating. for details about all available IOCP statements and keywords.

The PCHID 5A2 has been merged to the IOCDS by the CMT which assigned CHPID 7E to PCHID 5A2. * PARTITION=((CSS(0).5D.54.3F).LINK=((CSS(2).First we explain how the CHPIDs are defined.50.5C.* 51.(=)).54.59.0E)).2).256)).* 28.55.50. With the CHPIDs defined. LPAR names specified in IOCDS correspond to system names SC30 and SC31.(CSS(1).(CSS(3).28.A25.82)).A24.001)).0E).51.51.2C.(CSS(2).(A01).58.55.0E.59.0A).0A.82).TYPE=FC The PATH keyword in the CHPID statement defines CHPID 7E in CSS 2 as shared.28.54.55.2C.0E). Example 6-1 CHPID definition for switched point-to-point configuration CHPID PATH=(CSS(0. 120 FICON Planning and Implementation Guide .2C.(A23.* 2C.5D)).51.(CSS(3).TYPE=FC CHPID PATH=(CSS(1. * UNIT=2107 CNTLUNIT CUNUMBR=D100.(CSS(1).28.(CSS(1).0A).PCHID=5D2.58.0A.28.28.(A23. * UNITADD=((00. The same definition rules apply to the definition of CHPID 82.* (=))).7E).7E. D000.UNIT=2032 CNTLUNIT CUNUMBR=D000.82). The destination ports assigned for control units D000 and D100 are ports 28 and 3F in the FICON Director.(=)).2). * PATH=((CSS(0). * UNITADD=((00.5D). * LINK=((CSS(0).54.(CSS(3).82).0A.A29). See Example 6-1 for the CHPID statement and its keywords.0A)).7E. The LPAR names A23 and A24 are specified in the PARTITION keyword to use CSS 2 to access CHPID 7E.0E.SWITCH=65.SHARED. * UNITADD=((00.0A.SWITCH=65. Example 6-2 displays the CNTLUNIT statement and keywords. The SWITCH keyword is used to provide the switch number (65) of the Director to which the channel is connected.(CSS(2).(CSS(3).5C). * PARTITION=((CSS(1). and D100.28.FE.CUADD=0.(CSS(2).0A.(CSS(2).5C).CUADD=1.50.PATH=((CSS(2).59. Example 6-2 CNTLUNIT definition for switched point-to-point configuration CNTLUNIT CUNUMBR=0065. * LINK=((CSS(0).2C.* 50.59.(A11).256)).2C.A25.(CSS(1). * PATH=((CSS(0).0E.7E.5C)).58.SHARED.2C.3F).28.28. we next show how the CUs attached to the CHPIDs are defined.0E. The link address of control unit 0065 is specified as FE in the LINK keyword.FE)).PCHID=5A2.A24.28.* (=))).58. which is the port number of CUP in the FICON Director. * UNIT=2107 There are three control units defined by the CUNUMBER keyword: 0065.(CSS(2).2C.A29).55. The TYPE of the channel is specified as FC (native FICON).(CSS(2).5D).0E.

CHIPDs 7E and 82 have access to all CUs because both CHPIDs are defined in the PATH keyword on all three CUs.STADET=Y. Figure 6-3 on page 122 illustrates our configuration. For considerations regarding FCTCs.113).143).The logical CU image number 0 is specified for control unit D000. The same definitions are done for devices associated to control unit D100. Device 0065 belongs to control unit 0065. Chapter 6.STADET=Y. See Example 6-3 for the IODEVICE statement and its keywords.STADET=Y. The CHPIDs and the CUs are now defined. we have defined FICON Channel-to-Channel (FCTC) CUs and devices to allow data exchange between LPARs. “Adding FICON CTC connections” on page 305.STADET=Y.143). Defining FICON Channel-to-Channel (FCTC) In addition to storage control units. storage devices and the CUPs in the FICON Directors.CUNUMBR=(D000).UNIT=3390A The IODEVICE statement specifies the characteristics of devices 0065.UNIT=3390B IODEVICE ADDRESS=(D171. The next step is to define the devices owned by the CU.STADET=Y. In addition 143 alias Hyper PAV devices are defined starting at device address D071 and D171. Example 6-3 IODEVICE definition for switched point-to-point configuration IODEVICE ADDRESS=065.UNIT=3390A IODEVICE ADDRESS=(D100. The logical CU image number 1 is specified for control unit D100 in the CUADD keyword. refer to Appendix G. we are going to configure FCTCs in a switched point-to-point environment.CUNUMBR=(D100). and D100.CUNUMBR=(D000). starting at device address D000.UNITADD=00.CUNUMBR=(0065). * UNIT=2032 IODEVICE ADDRESS=(D000.113). which is the CUP port in FICON Director. Based on those considerations.CUNUMBR=(D100). Configuring a switched point-to-point topology 121 .UNIT=3390B IODEVICE ADDRESS=(D071. D000. These are base devices in the storage unit. On control unit D000 there are 113 devices are defined.

004)).UNIT=FCTC IODEVICE ADDRESS=(5034.PATH=((CSS(2). LINK=((CSS(2). Channels are attached to FICON Director 65 on ports 03 and 1B. LINK=((CSS(2).UNITADD=04.03)).UNITADD=((04.A23)). STADET=Y. Example 6-4 shows the definition of FCTC CUs and devices. Example 6-4 IOCDS definitions for FICON CTC CNTLUNIT CUNUMBR=5034.PARTITION=((CSS(2).UNIT=FCTC IODEVICE ADDRESS=(4044.A24)).UNIT=FCTC CNTLUNIT CUNUMBR=4044.PARTITION=((CSS(2).UNITADD=((04.UNIT=FCTC IODEVICE ADDRESS=(4034.03)). STADET=Y.PATH=((CSS(2).004). 122 FICON Planning and Implementation Guide .CUADD=24.7E)).1B)).004).UNIT=FCTC * * * * * * * * Four control units are defined by the CNTUNIT statement: 4044 and 5044 for direct access to LPAR A23. Control units 4034 and 5034 allow access to LPAR A24. there four logical control units defined: two of them for each partition for direct access in the same server. and 4034 and 5034 for LPAR A24. Because CHPIDs 7E and 82 are shared by logical partitions.UNIT=FCTC IODEVICE ADDRESS=(5044.PATH=((CSS(2). and two of them for each partition to access over the FICON link.A24)).CUNUMBR=(5044).7E)).A23)). both CHPIDs can be used by both partitions for FCTC communication.z10 EC SCZP201 LPAR SC30 A23 z/OS V1R10 FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E SAN768B Port 03 CU# CU# CU# CU# 4044 5044 4034 5034 LPAR SC31 A24 z/OS V1R10 FICON Express8 LX CSS2 PCHID 5D2 CHPID 82 Switch # 65 Switch @ 65 Port 1B Figure 6-3 FICON CTC (FCTC) configuration In this configuration there are two LPARs (A23 and A24) residing in the same z10 server.UNITADD=04.PATH=((CSS(2).82)).CUNUMBR=(4034).CUADD=23.82)).PARTITION=((CSS(2).004).UNITADD=04.UNITADD=((04.UNITADD=04.004)).UNITADD=((04.UNIT=FCTC CNTLUNIT CUNUMBR=5044.004)).CUNUMBR=(4044). Control units 4044 and 5044 allow direct access to LPAR A23.PARTITION=((CSS(2). Both LPARs can access CHPIDs 7E and 82 due to shared definition of those channel paths.004). STADET=Y.004)).CUNUMBR=(5034).1B)).CUADD=24.CUADD=23. For FCTC communication. LINK=((CSS(2). LINK=((CSS(2). STADET=Y.UNIT=FCTC CNTLUNIT CUNUMBR=4034.

This allows you to plan and define future I/O configurations and store them on the Support Element of the server. only two logical CUs are shown. Figure 6-5 on page 124 shows the Activate or Process Configuration Data menu. To simplify the illustration. Up to four IOCDS files can be stored in a System z server. Chapter 6. select option 2 (Activate or process configuration data). On the HCD start menu. Activating the changes (IOCDS) and configuring online After the definition of the desired configuration is finished. LPAR A23 Data transfer CU 4044 7E 03 @65 1B 82 CU 5034 Data transfer LPAR A24 Figure 6-4 Corresponding FCTC devices Data is transferred over the FICON link in both directions between the CUs and the logical partitions (LPARs). Figure 6-4 illustrates the logical view of the FCTC configuration and the data path between two corresponding FCTC devices. This allows data traffic to flow between both LPARs using the corresponding devices. The reverse applies to LPAR A24 when data is sent to or received from A23. as shown here: SC30 4044 4045 4046 4047 <----> <----> <----> <----> SC31 5034 5035 5036 5037 SC30 5044 5045 5046 5047 <----> <----> <----> <----> SC31 4034 4035 4036 4037 The rules we followed for device numbering are described in “FCTC device numbering scheme” on page 306. Refer to HCD User’s Guide. for detailed descriptions of all activation procedures. the data must be saved into the IODF data set and the IOCDS file on the Support Element.The PARTITON keyword specifies that devices in control unit 5034 and 4034 allow direct access only by LPAR A24. The tasks that are required to save definition data and activate the changes dynamically can be done via HCD. LPAR A23 sends data to A24 via CU 5034 and receives data from A24 via CU 4044. SC33-7988. The same rule applies to LPAR A24. Configuring a switched point-to-point topology 123 . These devices can be accessed by LPAR A23 over the FICON link by CHPIDs 7E and 82.

that a POR causes a system outage. The newly built IOCDS based on the I/O definition in the IODF should be sent to the CPCs Support Element and marked for use at the next POR. For this reason. where xx in the data set name is a suffix to identify the IODF among other IODFs. There is no Power-on Reset (POR) required because changes are directly updated to the HSA in the CPC. then we suggest that you plan ahead to make any changes in I/O definitions so that the new IOCDS is ready for use at the next POR. which may already exist. The suffix of the IODF data set is used in the IPL parameter when the z/OS image is brought up. The IOCDS is built using the definitions saved in the IODFxx data set. Keep in mind. This is the preferred method for activating changes. Using Activate or verify configuration dynamically (option 6) allows you to dynamically activate changes in the I/O definitions. Another way to activate changes in I/O definitions is to perform a POR at the CPC with the newly built IOCDS. This ensures that the desired IODF used by the operating system image matches the IOCDS that is loaded into the HSA. 124 FICON Planning and Implementation Guide . however. Build production I/O definition file Option 2. Activate or verify configuration dynamically The Build production I/O definition file (IODF) function (option 1) saves definition data into an IODFxx dataset.Figure 6-5 HCD Activate or Process Configuration Data menu The options on the menu to activate the changes are: Option 1. if a POR is required. Changes can be made at any time and saved in an IODF data set. Build IOCDS Option 3. Using Build IOCDS (option 2) allows you to build an IOCDS and save it to the CPC. This ensures that the newly built IOCDS will be used the next time the CPC is activated.

For planning and implementation information.CONFIG at the z/OS operator console. shown in Example 6-6. refer to IBM System Storage DS8000: Architecture and Implementation. “Planning the FICON environment” on page 59.CONFIG command IOS506I 09. enter D IOS. * SYSTEM=(2097.2009-04-21 16:12'. Configuring a switched point-to-point topology 125 .'09-04-21'.24 I/O CONFIG DATA 030 ACTIVE IODF DATA SET = SC31. 1 CHANNEL MEASUREMENT BLOCK FACILITY IS ACTIVE Example 6-5 shows that the IODF73 data set is currently active in z/OS image SC31. CSS2 is the active CSS in system SC31.* 00000000.LSYSTEM=SCZP201.MSG2='SC31. Configuring the FICON Director The FICON Director must be configured to match the designated configuration (shown in Figure 6-1 on page 116). SG24-6786. this includes: Specifying the Domain ID in the Director Enabling the ports Checking the port settings Other settings in the FICON Director are optional and should be specified after reading the considerations explained in Chapter 4.To verify which I/O definitions (IODF) are currently active. This proves that the IOCDS was built from the current active IODF data set and that the latest changes and new IO definitions were successfully activated. IODF73). “Configuring the DS8000 for FICON” on page 259 to configure the host adapter ports.00800006991E2094161224920109111F00000000. Compare the information provided by the D IOS. * TOK=('SCZP201'. It also provides information about the CSSs used in this z/OS image.IODF73 . The response to the command is shown in Example 6-5.08. Follow the procedure described in Appendix D. For the desired switched point-to-point topology. Chapter 6.IODF73 CONFIGURATION ID = TEST2097 EDT ID = 01 TOKEN: PROCESSOR DATE TIME DESCRIPTION SOURCE: SCZP201 09-04-21 16:12:24 SC31 IODF73 ACTIVE CSS: 2 SUBCHANNEL SETS CONFIGURED: 0. The configuration steps are based on the configuration shown in Figure 6-1 on page 116.'IODF73') The suffix of the IODF data set shown in the IOCDS ID statement and displayed by the D IOS command should be the same number (that is. Configuring the IBM Storage System DS8000 The host adapter ports in the DS8000 storage control unit have to be configured for native FICON protocol.'SC31'. Example 6-5 D IOS. Make sure that the selected ports (LX or SX) in the Director match the FICON node (LX or SX) that will be attached.CONFIG command with the data of the ID statement in the ICODS.1).'16:12:24'. Example 6-6 ID statement in IOCDS ID MSG1='IODF73'.

CU host adapters. “Configuring FICON Directors” on page 151.2 “Documentation” on page 62. when recabling or for problem determination). enter SETIOS ZHPF. so enter D M=CHP(7E) on the system console to display the current status of CHPID 7E. Configuring the CHPIDs online After you successfully activate the new I/O definitions and the FICON Director and the storage control unit are configured. Return to this section after the FICON Director is configured and ready for use. as shown in Example 6-8. as shown in Example 6-7). Refer to 4. For further information and considerations regarding fiber optic cabling and documentation.ONLINE IEE712I CONFIG PROCESSING COMPLETE To achieve the best performance on the FICON channel. the fiber optic cables have to be plugged into the FICON channels.23 ZHPF FACILITY 021 126 FICON Planning and Implementation Guide . Enter D IOS. to the FICON channel. Example 6-8 Enable zHPF D IOS. If the CHPID is offline.For LX type lasers. shown in Figure 6-1 on page 116. Ensure all fiber optic cables are labeled and documented for later use (for example. Note that even if all the definitions in the IOCDS are correct.11 “Physical connectivity” on page 93 and 4. This enables zHPF. and the FICON Director. enter CF CHP(7E). The channel status should change to ONLINE.9. If zHPF is disabled.1 “Configuration flowchart” on page 152 and follow the procedures described there. make sure that zHPF is enabled.YES.6 “High Performance FICON” on page 88 for considerations regarding how to exploit zHPF on System z servers. All fiber optic cables must have an LC duplex connector at both ends. Connecting the fiber optic cables After all definition and configuration tasks are completed.All the required steps to configure the FICON Director are described in detail in Chapter 8. refer to 4. Go to 8.27.ZHPF at the z/OS console to display zHPF settings for the system. Plug the fiber optic cable connectors to the designated ports.ZHPF RESPONSE=SC30 IOS630I 13. the CU.ON to configure the channel online. and the FICON Director. a 9µm single mode (SM) fiber optic cable is recommended to achieve maximum performance and distance. as listed here: PCHID 5A2 to FICON Director port 03 PCHID 5D2 to FICON Director port 1B CU host adapter port 0003 to FICON Director port 28 CU host adapter port 0242 to FICON Director port 3F Make sure the fiber optic cable ends are cleaned before plugging them into the transceiver. All the ports in our configuration are long wavelength (LX) laser types. the channel status might be offline. Example 6-7 CHPID 7E online state CF CHP(7E) IEE502I CHP(7E). you can configure the channel path online and try to communicate with the devices.1.

LOGICAL SWITCH ID = 65 ATTACHED ND = SLKWRM. Example 6-10 D M=DEV(0065. . However.CA. Information is also displayed about the attached node and the supported facilities (for example.DCX. Configuring a switched point-to-point topology 127 . In Example 6-10 we checked the status of device 0065.CA.BRD. ONLINE DEVICE STATUS FOR CHANNEL PATH 7E 0 1 2 3 4 5 6 7 8 9 A B C D E 0006 .1AFX0642C00Z. . . .0000 SCP DEVICE NED = SLKWRM.HIGH PERFORMANCE FICON FACILITY IS ENABLED The SETIOS ZHPF=YES command enables zHPF temporarily. The command output is shown in Example 6-9. To permanently enable zHPF for z/OS.DCX.0AFX0642C00Z PHYSICAL CHANNEL ID = 05A2 FACILITIES SUPPORTED = ZHPF F . and device D000 which is a storage device. + . .57 DISPLAY M 161 DEVICE 0065 STATUS=ONLINE CHP 7E 82 ENTRY LINK ADDRESS 03 1B DEST LINK ADDRESS FE FE PATH ONLINE Y Y CHP PHYSICALLY ONLINE Y Y PATH OPERATIONAL Y Y MANAGED N N CU NUMBER 0065 0065 MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = SLKWRM. Example 6-9 D M=CHP(7E) D M=CHP(7E) IEE174I 13. + + + HA HA HA Example 6-9 shows that CHPID 7E is online and working in a switched point-to-point configuration.PARMLIB member IECIOSxx to FCX=YES.22. . . zHPF). DESC=FICON SWITCHED.46 DISPLAY M 144 CHPID 7E: TYPE=1B.1AFX0642C00Z.BRD.002F SCP TOKEN NED = SLKWRM.CA. To check that communication to the attached devices is working properly.BRD. .DCX. after the next system IPL. It provides information about the channel and the attached devices.11. where xxxx is any device number. .DCX. which is the CUP in FICON Director. the zHPF facility will be reset to the default (disabled).1AFX0642C00Z.CA. 0D00 + + + + + + + + + + + + + + + 0D01 + + + + + + + + + + + + + + + 0D02 + + + + + + + + + + + + + + + … 0DFD HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFE HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFF HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA SWITCH DEVICE NUMBER = 0065 DEFINED ENTRY SWITCH .BRD. enter D M=DEV(xxxx).0000 Chapter 6.D000) IEE174I 14. . Now you can query the status and functional details of the channel path by entering D M=CHP(7E) at the operator console. change the FCX parameter in SYS1. . .

and the FICON Director are done.0000 SCP DEVICE NED = 002107. and to check the status of channels and devices. Information about the device (for example.0300 SCP TOKEN NED = 002107. ZHPF The response to D M=DEV(0065. check the status of PCHID 5A2 in the z10 server. the control unit number. Display the list of channels installed on the CPC.D000) displays all available paths to the devices and their status. the link address.0000000BALB1.IBM. the DS8000.DEVICE D000 STATUS=ONLINE CHP 50 54 58 5C 7E 82 ENTRY LINK ADDRESS 0F 0F 1B 1B 03 1B DEST LINK ADDRESS 2C 2C 0A 0A 28 3F PATH ONLINE N N Y Y Y Y CHP PHYSICALLY ONLINE N N Y Y Y Y PATH OPERATIONAL N N Y Y Y Y MANAGED N N N N N N CU NUMBER D000 D000 D000 D000 D000 D000 MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107. device type).900. the DS8000 storage. From any HMC where the CPC (server) is defined (for example. Search the list for the desired PCHID (5A2) and double-click the PCHID icon.900. and the FICON Director.IBM. SCZP201). Verifying the installation After all the configuration tasks in the z10 server. First.0000000BALB1. you can verify that the current configuration matches the desired configuration shown in Figure 6-1 on page 116. start a Single Object Operation (SOO) to that CPC. and the functions supported by the device (MIDAW and zHPF) is shown.0000000BALB1.75. Figure 6-6 PCHID details 128 FICON Planning and Implementation Guide .900.IBM.75. This will display the PCHID details panel shown in Figure 6-6.0000 HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW. This verification should be done to ensure that the fiber optic cables are correctly plugged to the designated ports in the z10 server.75.

Note: If any of the data displayed on the PCHID or CHPID detail panel does not match to the desired configuration. Similar information can be found on the CHPID Details panel.CHPID 2. Configuring a switched point-to-point topology 129 . The owning image of PCHID 5A2 is A23 and A24 (among other images). click the Channel Problem Determination button. you must correct the definitions in the IOCDS. PCHID type is FICON Express8. CHPIDs assigned to PCHID 5A2 are shared across images. select the channel list for an operAting system image from the Support Element workplace. This proves that PCHID 5A2 has CHPID 2. To display the CHPID details. On either the PCHID details panel or the CHPID details panel. see Figure 6-8 on page 130.7E) provided on the details window matches the designated configuration. CSS. This will display the Channel Problem Determination panel. shown in Figure 6-7. check that the FICON channels are connected to the correct FICON switch and port.7E assigned and images A23 and A24 can access the channel. Next.7E) is assigned to PCHID 5A2. Repeat this checking on other channels that were recently configured. where you can select which information you want to display. Chapter 6.7E (and 0. information for PCHID 5A2 (CHPID 2. Figure 6-7 CHPID details The CHPID Details panel provides similar information to the PCHID details panel.Essential information for PCHID 5A2 is shown on the PCHID Details panel: PCHID status is Operating. As you can see.

Figure 6-8 Channel Problem Determination panel Select Analyze channel information and click OK. as shown in Figure 6-9. Figure 6-9 Analyze Channel Information window 130 FICON Planning and Implementation Guide . The Analyze Channel Information window is displayed. which provides information about the node attached to the FICON channel.

as shown in Figure 6-10 on page 132. Chapter 6. The lower right side displays information about the attached node. The two digits in the middle represent the port number (03). the fiber optic cables may not be plugged correctly and must be checked. check that the path to the control unit image is initialized correctly and properly defined. then check the fiber optic cable link between the z10 server and the FICON Director to ensure that it is plugged correctly. Important: Make sure that the Node status for both nodes is displayed as Valid. If any other status is shown.Information about the nodes are shown at the bottom part of the window. select Analyze Serial Link Status and click OK. The World Wide Node Name (WWN) and the World Wide Port Name (WWP) are also shown for each port. On the Channel Problem Determination panel (shown in Figure 6-8 on page 130). For the attached node. which matches our desired configuration (see Figure 6-1 on page 116). or the Tag value and the WWNN or WWPN are not correct. examine the Tag field for each node. The Analyze Serial Link Status window is displayed. the Tag value represents the port number (03). If the displayed values are not as expected. The two left-most digits represent the switch address (65). The Tag provides information about the port number of the node. Next. check the Channel link address field. The two right-most digits of the Tag value represent the CHPID number for the channel node (7E). Next. The lower left side displays information about the node in the System z server. After completing the preceding steps and proving that the physical path to the FICON Director and the logical definitions of the link are correct. which displays the switch address and the port number of the attached Director. Check that the Type/model information and the serial number (Seq. In our scenario. If the node status is not Valid. number) are as expected. none of the displayed information is valid. we are now sure that PCHID 5A2 has connectivity to port number 03 on switch address 65. Configuring a switched point-to-point topology 131 . and may be used to prove that the channel is connected to the correct FICON Director if the WWN or WWP is known.

and that the fiber optic cable link between the Director and the CU has the correct cable type and plugging. CU address 00 and 01 correspond to control unit number D000 (CUADD=0) and D1000 (CUADD=1). you must check that the ports in the FICON Director and the CU are correctly configured. 132 FICON Planning and Implementation Guide . Link address FE is the CUP port in the FICON Director.Figure 6-10 Analyze Serial Link Status The Analyze Serial Link Status window provides status information about the link to the control unit images defined in the IOCDS.7E) is connected to port 03 on switch address 65 (switch number and channel link address). Link address 28 is the destination port where the CU is physically attached. which is used for communication and managing the FICON Director. The link to CU address 00 and 01 (among other CU images) is initialized. Figure 6-10 shows a link status of Initialization Complete for all defined CU images on link address FE and 28. Information is also displayed showing that PCHID 05A2 (CHPID 2. If the link status Initialization Complete is not shown. as defined in the IOCDS.

2005. Configuring a cascaded FICON Director topology This chapter describes all the tasks that are required to configure and define a FICON environment for a cascaded FICON Director topology. All rights reserved. Establishing a cascaded FICON Director topology involves the following: Defining FICON channel. 2009. 2006.7 Chapter 7. control unit. and device definitions in IOCP Activating the IOCDS Configuring the storage control unit Varying the channel path and device online Verifying the configuration © Copyright IBM Corp. 133 .

which correspond to the LPAR names A23 and A24. with the following items physically installed: System z server DS8000 storage system FICON Directors Fiber cabling infrastructure 7. Both LPARs have access to the two FICON Express8 channels. The channel adapters are FICON 134 FICON Planning and Implementation Guide . The system names running in this partition are SC30 and SC31. In our example case.7.1. The operating system that is running in both partitions is z/OS V1R10. including setting up FICON Directors and configuring the attached storage controller. consisting of a z10 EC server. Our configuration is shown Figure 7-1. “Planning the FICON environment” on page 59. the entire infrastructure is already installed in our environment. Two FICON Express8 channels are defined as shared and assigned to CSS2. z10 EC SCZP201 LPAR SC30 A23 z/OS V1R10 FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E System Storage SAN384B Switch # 66 Switch @ 66 Port 8D Port C4 Port 88 Port CF DS8000 CU# D000 LX 0003 D0xx LPAR SC31 A24 z/OS V1R10 FICON Express8 LX CSS2 PCHID 5D2 CHPID 82 Port 03 Port 2D Port 34 Switch # 65 Switch @ 65 LX 0242 D1xx CU# D100 Port 1B * All cable connectors are LC Duplex type SAN768B Figure 7-1 Cascaded FICON configuration The z10 EC server (SCZP201) has two LPARS defined and activated. which will be used for data transfer to the DS8000 storage device. Our discussion is based on an infrastructure that is already built and installed.1 Description of our environment Based on the considerations explained in Chapter 4. and a DS8000 storage device. the z10 server supports zHPF protocol. a cascaded FICON configuration will be built.1 Establishing a cascaded FICON Director topology This section describes the tasks required to configure a cascaded FICON Director topology on a System z server. Longwave laser optic and SM fiber optic cables are used to achieve maximum distance and performance. a SAN768B and SAN384B FICON Director. As a standard feature.

Tasks Figure 7-2 on page 136 shows the main steps required to define and activate a cascaded FICON Director configuration. The switch number (Switch #) and the switch address (Switch @) are both set to 66. which is required to support cascaded FICON Directors. The SAN384B FICON Director has connectivity to storage devices on port 88 and CF. as well as to the ports in the FICON Directors and to the DS8000 host adapters. The SAN768B FICON Director has connectivity to the z10 server on port 03 and port 1B. The DS8000 will have the zHPF feature enabled to communicate to the z10 server. If the Directors are connected they build a High Integrity Fabric. All the ports in the FICON Director are longwave (LX) ports. A maximum unrepeated distance of 10km (6. 7. The fiber optic cables have an LC Duplex connector at both ends to connect to the z10 FICON Express8 channels. PCHID 5D2 has CHPID 82 assigned. The switch address is the Domain ID specified in the Director.2 miles) is supported by the longwave laser (LX) feature when using 9µm single mode (SM) fiber optic cables. PCHID 5A2 has CHPID 7E assigned.1. The switch number (Switch #) and the switch address (Switch @) are both set to 65. which consists of two fiber optic cables connected to ports 2D and 34 in switch 65 and ports 8D and 34 in switch 66. The switch number is specified in the IOCDS. The ISLs are transparent in the path from the System z server to the CU and do not require any definitions in IOCDS. An IBM System Storage DS8000 will be connected to the SAN384B FICON Director on port 88 and the CF. The switch number is specified in the IOCDS. The two host adapters are installed in the DS8000 supporting longwave (LX) laser. using a longwave laser for data traffic through the fiber optic cable. The switch address is the Domain ID specified in the Director. The FICON Directors are attached by an Inter-Switch Link (ISL). Configuring a cascaded FICON Director topology 135 . Chapter 7.2 Tasks and checklist This section provides an overview of required tasks.Express8 LX features (FC 3325). Port numbers 0003 and 0242 in the DS8000 are used to connect to the FICON Director. Two logical control units (D000 and D100) are defined which have devices D000-D0FF or D1000-D1FF assigned. as well as a checklist that we used to make sure that all hardware and software prerequisites were met before starting configuration tasks. All the ports in the FICON Director are longwave (LX) ports.

“Configuration flowchart” on page 152.Verification checklist Follow the verification checklist to ensure that all hardware and software prerequisites are met. Information about how to verify that your actual configuration matches the desired configuration is given in “Verifying the installation” on page 146.1. Check that FICON features are available on the System z server to establish the desired configuration. CU. Both hardware and software requirements must be checked. FC 3324 SX FC 3318. Configure FICON Director Plug fiber cables Information about fiber optic cables and plugging rules to achieve the desired configuration is given in “Connecting the fiber optic cables” on page 144. and storage device in IOCDS” on page 138. Information about defining the channel paths. z/OS V1. FC 3322 FICON Express8 LX FC 3325 SX FC 3326 If using the FICON Express8 feature. FC 3323. FICON Express2 LX FC 3319 SX FC 3320 FICON Express4 LX FC 3321. Verify configuration Figure 7-2 Main steps for configuring and verifying a cascaded FICON Director configuration Verification checklist Before configuring the cascaded FICON topology shown in Figure 7-1 on page 134. 136 FICON Planning and Implementation Guide . the following list was checked. and devices on a System z server Configure storage CU The configuration tasks for a DS8000 storage system are described in “Configuring the IBM Storage System DS8000” on page 143. control units. All steps in the checklist must be finished and corrected (if required) to ensure a smooth and successful configuration of the topology. CU. All required steps to set up the FICON Directors are shown in 8.7 with the IBM Lifecycle Extension for z/OS V1. check the System z operating system requirements.1. Go to “Verification checklist” on page 136.7 (5637-A01) is required to support the FICON Express8 feature in a z10 server. and devices is given in “Defining channel. Define channel. see “System z FICON feature support” on page 25. For details about each feature code.

save.2 miles) at maximum speed. LX SFPs SX SFPs Check that high-integrity fabric support is enabled on the FICON Directors. as well as the ports in the FICON Director and the DS8000 host adapter.1. We used HCD to create. 7. An LC duplex connector is required at both ends of the fiber optic cables to connect to the System z server FICON adapter. Check that the correct fiber optic cables are available to connect the FICON Director to the z10 server and the DS8000 storage to the FICON Director. Check the DS8000 storage hardware requirements to support FICON longwave (LX) or shortwave (SX) connectivity.5µm multi mode (MM) fiber optic cable is required to support shortwave laser (SX). Chapter 7.Check the 2097DEVICE PSP bucket for the latest information about FICON Express support on the operating system. CU. Check that the zHPF feature is enabled in the DS8000. The steps include: Defining channel. and activate our I/O definitions. Note: All fiber optic cables used in a link must be of the same type.16. FICON Director. and storage device in IOCDS Activating changes (IOCDS) and configuring online Configuring the IBM Storage System DS8000 Configuring the FICON Directors Connecting the fiber optic cables Configuring the CHPIDs online Verifying the installation For our scenario. A 9µm single mode (SM) fiber optic cable is required to support longwave laser (LX) for a maximum distance of 10km (6. A 50µm or 62. they must be either all single mode or all multi-mode fiber optic cables. and the storage device online and operating. Check that FC 0709 and FC 7092 are installed to support zHPF. See “System z FICON feature support” on page 25 for the maximum supported distance.3 Getting started This section describes all the tasks that are required to achieve the designated configuration shown in Figure 7-1 on page 134.0 or higher to support the zHPF feature. Note: FC 0709 and FC 7092 must be ordered to obtain the license key to enable zHPF support in a DS8000 storage controller. depending on cable type and speed. Configuring a cascaded FICON Director topology 137 .1. For example. It explains what needs to be done to get the FICON channels. storage control units. Check that the DS8000 firmware is at level 6. Check that the number and types of FICON Director ports match the configuration requirements. we had an active partition in a System z server running z/OS with HCD.

TYPE=FC The PATH keyword in the CHPID statement defines CHPID 7E in CSS 2 as SHARED. for details about all IOCP statements and keywords.(A23.(CSS(2). Refer to HCD User’s Guide.A24. Because no system outage is required for dynamic changes. * PARTITION=((CSS(1). CUs.(A01).A29). the changes are directly updated into the Hardware System Area (HSA) of the server. The Support Element in the server stores the IOCDS on its built-in HDD. First.TYPE=FC CHPID PATH=(CSS(1. PCHID 5A2 has been merged to the IOCDS by the CMT. SB10-7037.SHARED.(A11). SC33-7988.A29).PCHID=5A2.7E). Refer to IOCP User’s Guide.SWITCH=65. CNTLUNIT. this is the preferred method for changing I/O definitions.2).2). 138 FICON Planning and Implementation Guide . and storage device in IOCDS There are several tools available that you can use to define and configure a FICON environment on a System z server. for detailed information regarding the use of HCD. The newly created IOCDS will be sent to the Support Element of the desired System z server. The following examples show the CHPID.PCHID=5D2. The operating system uses the HSA to manage the I/O subsystem.(CSS(2). The SWITCH keyword is used to provide the switch number (65) of the Director to the channel to which it is connected. and IODEVICE statements if all the definition tasks based on the configuration shown in Figure 7-1 on page 134 are finished. HCD was used to define the CHPIDs. If changes to the I/O definitions are performed dynamically. CUs. Example 7-1 displays the CHPID statement and its keywords.Defining channel. the IOCDS has to be updated to ensure consistency between the software and the microcode in the hardware. Only statements and keywords relevant to our desired configuration are explained. The LPAR names A23 and A24 are specified in the PARTITION keyword to use CSS 2 to access CHPID 7E. After all definition tasks are completed. or devices can be performed dynamically. The TYPE of the channel is specified as native FICON (FC). CU. which assigned CHPID 7E to PCHID 5A2.(=)). “Configuration and definition tools” on page 255. the definition data is stored in the I/O definition file (IODF) data set.(A23. Example 7-1 CHPID definition for cascaded FICON configuration CHPID PATH=(CSS(0.* (=))). and devices for the sample configuration we describe in this section.A24. Changes or additions of channel paths. * PARTITION=((CSS(0).A25. The IOCDS stored in the Support Element is loaded into System z storage during a Power-on Reset (POR) of the CPC. The HSA is a separate storage area in the System z server that holds all I/O specifications and provides status information for all defined I/Os.SWITCH=65.82).(=)). The LPAR names specified in ICODS correspond to system names SC30 and SC31.A25. Any time you make changes to the I/O configuration. we explain how the CHPIDs are defined.* (=))). The definition tools and an explanation of the statements and keywords used in the IOCP are described in Appendix C.SHARED.

The LINK keyword defines the destination port address the control unit to which it is attached.65FE)).CUNUMBR=(0066).7E.* 2C.2C.LINK=((CSS(2).UNIT=3390A Chapter 7. 65FE and 66FE).STADET=Y.51.7E.STADET=Y.82).7E.0A). as specified in the PATH keyword for all four CUs.66FE.66FE)).(CSS(2). * LINK=((CSS(0).82).256)). Example 7-2 displays the CNTLUNIT statement and keywords.54. * LINK=((CSS(0).(CSS(3).0E). we illustrate next how the CUs that are attached to the CHPIDs are defined.0A. The destination ports specified for control units D000 and D100 are 6688 and 66CF.5C).STADET=Y. * UNIT=2032 IODEVICE ADDRESS=(D000.82)).6688.58.2C.5D). Example 7-2 CNTLUNIT definition for a cascaded FICON configuration CNTLUNIT CUNUMBR=0065. * UNITADD=((00.0E.55.66CF).0A.(CSS(2).143).CUNUMBR=(D000).CUNUMBR=(0065).0E.50.5D)).CUNUMBR=(D100).143). Example 7-3 IODEVICE statement for cascaded FICON configuration IODEVICE ADDRESS=065.2C.82)). With the CHPIDs defined.58.50.0A.55.28. 0066.59.51.001)).7E.PATH=((CSS(2).55. * UNIT=2107 CNTLUNIT CUNUMBR=D100.UNITADD=00.28. Control units 0065 and 0066 are the CUP ports (port number FE) in the FICON Directors defined on the LINK statement (in our example.* 50.UNIT=3390A IODEVICE ADDRESS=(D100.STADET=Y.CUADD=1.STADET=Y.001)). The first byte of the link address represents the switch address and the second byte defines the port address in that switch. The CHPIDs and the CUs are now defined.59.28.The same definition rules apply to the definition of CHPID 82.50.0A.(CSS(1).UNIT=2032 CNTLUNIT CUNUMBR=D000.(CSS(3). * UNITADD=((00.54.(CSS(1).LINK=((CSS(2). The logical CU image number 0 is specified for control unit D000.UNIT=2032 CNTLUNIT CUNUMBR=0066.0A.65FE.55.28.UNIT=3390B IODEVICE ADDRESS=(D071.28.UNIT=3390B IODEVICE ADDRESS=(D171. D000. * UNITADD=((00. * UNIT=2107 There are four control units defined by the CUNUMBR keyword: 0065.0E.0A)).5D).256)).5C)).* 28.(CSS(1).54.51. * UNIT=2032 IODEVICE ADDRESS=066. * PATH=((CSS(0).113).54.66CF).CUNUMBR=(D100).59.0A). The next step is to define the devices owned by the CU.59.28.58.(CSS(3).2C.28.0E.2C.CUADD=0.5C.(CSS(2).58.(CSS(2).* 51.STADET=Y.(CSS(3). and D100. * PATH=((CSS(0). Example 7-3 displays the IODEVICE statement and keywords.PATH=((CSS(2). CHPIDs 7E and 82 in CSS2 have access to all control units.0E)).(CSS(1).5D.UNITADD=00.5C).6688.0E). Configuring a cascaded FICON Director topology 139 . * UNITADD=((00.2C.0E.2C. The logical CU image number 1 for control unit D100 is specified in the CUADD keyword.CUNUMBR=(D000).113).

and the CUPs in the FICON Directors. Figure 7-3 illustrates the desired FCTC configuration. Example 7-4 FCTC configuration CNTLUNIT CUNUMBR=5034.66CF)).004)). 143 alias devices are defined starting at device address D071 and D171. which is accessed by LPAR A23. although more than two control units could be defined. LPAR A24 will communicate with CU 4044. which is accessed by LPAR A24.PATH=((CSS(2). Each LPAR will have access to a logical control unit and their logical devices defined in the corresponding LPAR in different z10 servers. These are base devices in the storage unit. LINK=((CSS(2). starting at device address D000. “Adding FICON CTC connections” on page 305.CUADD=23. Example 7-4 shows the definition of FCTC control units and devices. Defining a FICON Channel-to-Channel (FCTC) connection In addition to storage control units.7E)). To simplify the illustration. Device 0065 belongs to control unit 0065 which is the CUP port in FICON Director. we are going to configure FCTCs in a cascaded FICON environment.CUNUMBR=(5034). and CHPID 82 to port CF in switch 66.UNIT=FCTC IODEVICE ADDRESS=(5034.The IODEVICE statement specifies the characteristics of devices 0065. STADET=Y.PARTITION=((CSS(2). For considerations regarding FCTCs.004). On control unit D000 there are 113 devices defined. Based on those considerations. The same definitions are done for devices associated to control unit D100. D000. refer to Appendix G. LPAR A23 will communicate with CU 5034. two control units and two LPARs are shown in this sample confirmation. z10 EC SCZP201 LPAR SC30 (A23) z/OS V1R10 FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E System Storage SAN384B Switch @ 66 Switch @ 66 Port 8D Port C4 Port CF z10 EC SCZP202 ISL Port 03 Port 2D Port 34 FICON Express8 LX PCHID 5D2 CHPID 82 CU# 5034 LPAR SC31 (A24) z/OS V1R10 CU# 4034 4044 CU# 4044 Switch # 65 Switch @# 65 CSS2 SAN768B Figure 7-3 FCTC configuration Both z10 servers are attached to cascaded FICON Directors: CHPID 7E to port 03 in switch 65. and D100.A24)). we have defined FICON Channel-to-Channel (FCTC) CUs and devices to allow data exchange between LPARs. In addition.UNITADD=04.UNITADD=((04.UNIT=FCTC 140 FICON Planning and Implementation Guide * * . storage devices.

The LINK keyword specifies the destination port to communicate to the control unit in the other server.UNITADD=((04.UNIT=FCTC IODEVICE ADDRESS=(4044. The reverse applies to LPAR A24 when data is sent to or received from A23.004). LINK=((CSS(2).UNIT=FCTC * * One control unit is defined for each LPAR: CU 5034 is defined for LPAR A23 and CU 4044 is defined for LPAR A24. Configuring a cascaded FICON Director topology 141 .PATH=((CSS(2). between the CUs and the logical partitions (LPARs). in both directions. CU 4044 resides in server SCZP201.PARTITION=((CSS(2).6503)). To simplify the illustration. LPAR A23 Data transfer CU 4044 7E 03 @65 ISL @66 CF 82 CU 5034 Data transfer LPAR A24 Figure 7-4 FCTC data transfer Data is transferred over the FICON link.82)). Up to four IOCDS files can be stored in a System z server.CNTLUNIT CUNUMBR=4044. This allows data traffic to flow between the LPARs using the corresponding devices. This allows you to plan and define future I/O configurations and store them on the Support Element of the server. Figure 7-4 illustrates the logical view of the FCTC configuration and the data path between two LPARs.UNITADD=04.A23)).004)).CUNUMBR=(4044). which is port 66CF for LPAR 23 to communicate with CU 5034. CU 5034 resides in server SCZP202. STADET=Y. Chapter 7. LPAR A23 sends data to A24 via CU 5034 and receives data from A24 via CU 4044. only two logical CUs are shown. and port 6503 for LPAR A24 with communicate to CU 4044. The PARTITION keyword in the IODEVICE statement specifies the logical partition that can access the device.CUADD=24. as follows: SC30 4044 4045 4046 4047 <---> <---> <---> <---> SC31 5034 5035 5036 5037 The rules we followed for device numbering are described in “FCTC device numbering scheme” on page 306. Activating changes (IOCDS) and configuring online When the definition of the desired configuration is finished. the data must be saved into the IODF data set and the IOCDS file on the Support Element.

Activate or verify configuration dynamically The Build production I/O definition file (IODF) function (option 1) saves definition data into an IODFxx dataset. for detailed descriptions of all the activation procedures.Keep in mind. that a POR causes a system outage. The suffix of the IODF data set is used in the IPL parameter when the z/OS image is brought up. For this reason. The IOCDS is built using the definitions saved in the IODFxx data set. Using Build IOCDS (option 2) allows you to build an IOCDS and save it to the CPC. where xx in the data set name is a suffix to identify the IODF among other IODFs which may already exist. Refer to HCD User’s Guide. There is no Power-on Reset (POR) is required because changes are directly updated to the HSA in the CPC. Figure 7-5 HCD Activate or process configuration data menu The options on the menu to activate the changes are: Option 1. On the HCD start menu. This ensures that the desired IODF used by the operating system image matches the IOCDS that is loaded into the HSA. This is the preferred method of activating changes. select option 2 (Activate or process configuration data). Build production I/O definition file Option 2. SC33-7988. Build ICODS Option 6. Using Activate or verify configuration dynamically (option 6) allows you to dynamically activate changes in the I/O definitions. Another way to activate changes in I/O definitions is to perform a POR at the CPC with the newly-built IOCDS. if a POR is required. Changes can be 142 FICON Planning and Implementation Guide . however. Figure 7-5 shows the Activate or process configuration data menu. then we suggest that you plan ahead to make any changes in I/O definitions so that the new IOCDS is ready for use at the next POR.The tasks that are required to save definition data and activate the changes dynamically are performed via HCD.

For the desired cascaded FICON Directors topology. The response to the command is shown in Example 7-5. It also provides information about the CSSs used in this z/OS image.MSG2='SC30. 1 CHANNEL MEASUREMENT BLOCK FACILITY IS ACTIVE Example 7-4 on page 140 shows that IODF78 data set is currently active in z/OS image SC30.2009-05-07 09:26'.28 I/O CONFIG DATA 003 ACTIVE IODF DATA SET = SC30. * TOK=('SCZP201'.'09-05-07'. * SYSTEM=(2097.1).* 00000000. Chapter 7. Configuring the IBM Storage System DS8000 The host adapter ports in the DS8000 storage control unit have to be configured for native FICON protocol.IODF78 . The newly built IOCDS based on the I/O definition in the IODF should be sent to the CPCs Support Element and marked for use at the next POR.'IODF78') The suffix of the IODF data set that is shown in the IOCDS ID statement and displayed by the D IOS command should be the same number (that is. shown in Figure 7-1 on page 134. This proves that the IOCDS was built from the current active IODF data set and that the latest changes and new IO definitions are successfully activated.CONFIG at the z/OS operator console. SG24-6786.'09:26:49'.CONFIG command IOS506I 13.IODF78 CONFIGURATION ID = TEST2094 EDT ID = 01 TOKEN: PROCESSOR DATE TIME DESCRIPTION SOURCE: SCZP201 09-05-07 09:26:49 SC30 IODF78 ACTIVE CSS: 2 SUBCHANNEL SETS CONFIGURED: 0. Example 7-6 ID statement in IOCDS ID MSG1='IODF78'. this includes: Specifying the Domain ID in the Directors Enabling the ports Checking the port settings Make sure that the selected ports (LX/SX) in the Director match the FICON node (LX/SX) that will be attached. to configure the host adapter ports.'SC30'. Compare the information provided by the D IOS. To verify which I/O definitions (IODF) are currently active. IODF78).CONFIG command with the data of the ID statement in the IOCDS.00800006991E2094092649000109127F00000000. as shown in Example 7-6. CSS2 is the active CSS in system SC30. Follow the procedure described in Appendix D. The configuration steps are based on the configuration shown in Figure 7-1 on page 134. “Configuring the DS8000 for FICON” on page 259. Configuring a cascaded FICON Director topology 143 . enter the command D IOS.10.made at any time and saved in an IODF data set. This ensures that the newly built IOCDS will be used the next time the CPC is activated. Example 7-5 D IOS. Configuring the FICON Directors The FICON Directors must be configured to match the designated configuration.LSYSTEM=SCZP201. For planning and implementation information refer to IBM System Storage DS8000: Architecture and Implementation.

fiber optic cables have to be plugged in to the FICON channels. a 9µm single mode (SM) fiber optic cable is recommended to achieve maximum performance and distance. when recabling or for problem determination). as listed here: PCHID 5A2 to FICON Director 65 port 03 PCHID 5D2 to FICON Director 66 port 1B CU host adapter port 0003 to FICON Director port 28 CU host adapter port 0242 to FICON Director port 3F The fiber optic cables between both FICON Directors (ISL) have to be plugged when you configure the FICON Directors. Example 7-7 CHPID 7E online state CF CHP(7E) IEE502I CHP(7E).All the required steps to configure the FICON Directors are described in detail in Chapter 8. make sure that zHPF is enabled. “Physical connectivity” on page 93 and 4. refer to 4. See 4.11. Plug the fiber optic cable connectors to the designated ports (shown in Figure 7-1 on page 134) to the FICON channel. the CU. however. “Configuring FICON Directors” on page 151.9.YES.ONLINE IEE712I CONFIG PROCESSING COMPLETE To achieve the best performance on the FICON channel. the channel path can be configured online and checked to see if we can communicate to the devices. This enables zHPF.ON on the z/OS operator console to configure the channel online.6. as shown in Example 7-7. CU host adapters. “Documentation” on page 62. Return to this section after the FICON Directors are configured and ready for use. “Configuration flowchart” on page 152 and follow the procedures described there. Connecting the fiber optic cables After all the definition and configuration tasks are completed. If the CHPID is offline. All fiber optic cables must have an LC duplex connector at both ends. Go to 8. Enter D IOS. the channel status might be offline. enter CF CHP(7E). Enter D M=CHP(7E) on the system console to display the current status of CHPID 7E. Note. “High Performance FICON” on page 88 for considerations about exploiting zHPF on System z servers.1.2. Make sure the fiber optic cable ends are cleaned before plugging them into the transceiver. If zHPF is disabled. Ensure that all fiber optic cables are labeled and documented for later use (for example. as shown in Example 7-8. The channel status should change to ONLINE. Configuring the CHPIDs online After the new I/O definitions are successfully activated. and from the CU host adapters to the FCION Director. enter SETIOS ZHPF. use the longwave (LX) type of laser. For the LX type of laser.1. 144 FICON Planning and Implementation Guide . and the FICON Directors. For further information and considerations regarding fiber cabling and documentation. All fiber links from the FICON channel to the FICON Director. and the FICON Director.ZHPF at the z/OS console to display zHPF settings for the z/OS image. that even if all definitions in the IOCDS are correct.

27. zHPF). .46 DISPLAY M 144 CHPID 7E: TYPE=1B. after the next system IPL. enter D M=DEV(0065. change the FCX parameter in SYS1. + .BRD. .11 DISPLAY M 986 DEVICE 0065 STATUS=ONLINE CHP 7E 82 ENTRY LINK ADDRESS 6503 651B DEST LINK ADDRESS 65FE 65FE PATH ONLINE Y Y CHP PHYSICALLY ONLINE Y Y PATH OPERATIONAL Y Y MANAGED N N CU NUMBER 0065 0065 Chapter 7. However. The command output is shown in Example 7-9. To permanently enable zHPF for z/OS.D000) D M=DEV(0065. . enter the command D M=DEV(xxxx). ONLINE DEVICE STATUS FOR CHANNEL PATH 7E 0 1 2 3 4 5 6 7 8 9 A B C D E 0006 . Information is also displayed about the attached devices and the supported facilities (for example. Example 7-10 D M=DEV(0065. . . It provides information about the channel and the attached devices. . For example. the zHPF facility will be reset to the default (disabled). + + + HA HA HA Example 7-9 shows that CHPID 7E is online and operating. Example 7-9 D M=CHP(7E) D M=CHP(7E) IEE174I 13. to check the status of device 0065 (the CUP in FICON Director 65) and device D000 (a storage device).D000) on a z/OS console.D000) IEE174I 11.CA.59.ZHPF RESPONSE=SC30 IOS630I 13.23 ZHPF FACILITY 021 HIGH PERFORMANCE FICON FACILITY IS ENABLED Using the SETIOS ZHPF=YES command enables zHPF temporarily. . . .DCX. Configuring a cascaded FICON Director topology 145 .0AFX0642C00Z PHYSICAL CHANNEL ID = 05A2 FACILITIES SUPPORTED = ZHPF F . To check that communication to the attached devices is working properly.22. 0D00 + + + + + + + + + + + + + + + 0D01 + + + + + + + + + + + + + + + 0D02 + + + + + + + + + + + + + + + … 0DFD HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFE HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFF HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA SWITCH DEVICE NUMBER = NONE ATTACHED ND = SLKWRM. as shown in Example 7-10.PARMLIB member IECIOSxx to FCX=YES. DESC=FICON SWITCHED. .Example 7-8 Enabling zHPF D IOS. . . Now you can query the status and functional details of the channel path by entering D M=CHP(7E) at the operator console. where xxxx is any device number.

0000000BALB1. Display the list of channels installed on the CPC.0AFX0642C00Z. MIDAW and zHPF).0000 DEVICE D000 STATUS=ONLINE CHP 50 54 58 5C 7E 82 ENTRY LINK ADDRESS 0F 0F 1B 1B 6503 651B DEST LINK ADDRESS 2C 2C 0A 0A 6688 66CF PATH ONLINE N N Y Y Y Y CHP PHYSICALLY ONLINE N N Y Y Y Y PATH OPERATIONAL N N Y Y Y Y MANAGED N N N N N N CU NUMBER D000 D000 D000 D000 D000 D000 MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107. start a Single Object Operation (SOO) to that CPC.0000 SCP DEVICE NED = SLKWRM.DCX. Search the list for the desired PCHID (5A2) and double-click the PCHID icon.DCX.900. device type). and to check the status of channels and devices.0AFX0642C00Z.CA.0000 SCP DEVICE NED = 002107.75. and the FICON Directors are completed. From any HMC where the CPC (server) is defined (for example.DCX.900. Information is provided about the device (for example. Verifying the installation After all the configuration tasks in the z10 server.BRD.BRD.0000000BALB1. and the functions supported by the device (for example. This verification should be done to ensure that the fiber optic cables are correctly plugged to the designated ports in the z10 server.BRD. the link address.IBM.IBM.0000000BALB1. ZHPF The response to D M=DEV(0065. the DS8000 storage.MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = SLKWRM.CA.IBM. This will display the PCHID details panel shown in Figure 7-6 on page 147.75.CA. First.0000 HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW. you can verify that the current configuration matches the desired configuration shown in Figure 7-1 on page 134. SCZP201).900. 146 FICON Planning and Implementation Guide .0AFX0642C00Z. check the status of PCHID 5A2 in the z10 server. the DS8000 and the FICON Directors. the control unit number.0003 SCP TOKEN NED = SLKWRM.0300 SCP TOKEN NED = 002107.75.D000) displays all available paths to the devices and their status.

Figure 7-6 PCHID details Essential information for PCHID 5A2 is shown on the PCHID Details panel: The PCHID status is Operating. Configuring a cascaded FICON Director topology 147 . Similar information can be found on the CHPID Details panel (see Figure 7-7). select the channel list for an operating system image from the Support Element workplace. As shown. The owning image of PCHID 5A2 is A23 and A24 (among other images). This proves that PCHID 5A2 has CHPID 2. The PCHID type is FICON Express8.7E) provided on the details window matches the designated configuration.7E) is assigned to PCHID 5A2. CSS. Repeat this checking on other channels that have been recently configured.7E (and 0. To display the CHPID details. Chapter 7. information for PCHID 5A2 (CHPID 2.7E assigned and images A23 and A24 can access the channel. The CHPIDs assigned to PCHID 5A2 are shared across images. Figure 7-7 CHPID details The CHPID Details panel provides similar information to the PCHID details panel.CHPID 2.

On either the PCHID details or CHPID details panel. Figure 7-8 Channel Problem Determination panel Select Analyze channel information and click OK. 148 FICON Planning and Implementation Guide . click the Channel Problem Determination button. you must correct the definitions in the IOCDS. The Analyze Channel Information window is displayed as shown in Figure 7-9 on page 149. where you can select which information you want to display. check that the FICON channels are connected to the correct FICON switch and port. and it provides information about the node attached to the FICON channel. Next. see Figure 7-8. This will display the Channel Problem Determination panel.Important: If any of the data displayed on the PCHID or CHPID detail panel does not match the desired configuration.

Be aware. Check that the Type/model information. is as expected. If any other status is shown. none of the information displayed is valid.Figure 7-9 Analyze Channel Information Information about the nodes is displayed at the bottom part of the window. If the node status is not Valid or the tag value and WWPN are not correct. The two right-most digits of the tag value represent the CHPID number for the channel node (7E). as well as the serial number (Seq. and may be used to prove that the channel is connected to the correct FICON Director (if the WWN or WWP of the attached device is known). Chapter 7. examine the tag field for each node. Next. The lower right side displays information about the attached node. depending on the vendor. and for the attached node it represents the port number (0003). The tag provides information about the port number of the node. you must check the fiber optic cable link between the z10 server and the FICON Director to ensure that it is plugged correctly. The World Wide Node Name (WWN) and the World Wide Port Name (WWP) are also shown for each port. number). Configuring a cascaded FICON Director topology 149 . however. The lower left side displays information about the node in the System z server. Important: Verify that the Node status shown for both nodes is Valid. that the tag value is provided by the attached device during link initialization and may have different meanings.

150 FICON Planning and Implementation Guide . The Analyze Serial Link Status window is displayed. The two digits in the middle represent the port number (03). Figure 7-10 shows a link status of Initialization Complete for all defined CU images on link address 65FE and 6628. In a cascaded FICON configuration. which matches our desired configuration (see Figure 7-1 on page 134). CU address 00 and 01 correspond to control unit number D000 (CUADD=0) and D1000 (CUADD=1). select Analyze Serial Link Status and click OK. Scroll through the list of CU images and check that status for all CUs is Initialization Complete. In this case. If the link status Initialization Complete is not shown. That means link address 65FE is the CUP port in switch 65. as defined in the IOCDS. The link to CU address 00 and 01 (among other CU images) is initialized. check the Channel link address field. The two left-most digits represent the switch address (65). shown in Figure 7-8 on page 148. check that the path to the control unit image is initialized correctly and properly defined. Link address 6688 is the destination port where the CU is physically attached. as shown in Figure 7-10. which shows the switch address and the port number of the attached Director. it means that the fiber optic cables may not be plugged correctly and must be checked. Figure 7-10 Analyze Serial Link Status The Analyze Serial Link Status window provides status information about the link to the control unit images defined in the IOCDS.7E) is connected to port 03 on switch address 65 (Switch number and Channel link address). If the displayed values are not as expected. In our scenario we are now sure that PCHID 5A2 has connectivity to port number 03 on switch address 65. After completing the preceding steps and proving that the physical path to the FICON Director and the logical definitions of the link are correct. and link address 6688 points to a storage control unit attached to switch 66. Information is also displayed that PCHID 05A2 (CHPID 2. and that the fiber optic cable link between the Director and the CU has the correct cable type and plugging.Next. On the Channel Problem Determination panel. you must check that the ports in the FICON Director and CU are correctly configured. you must use a two-byte link address to specify the link address.

It provides configuration examples for switched point-to-point and cascaded FICON Director topologies. Configuring FICON Directors This chapter explains how to configure IBM 2499 FICON Directors. refer to Cisco FICON Basic Implementation. 2009. 151 . 2006. All rights reserved. The following topics are discussed: Configuration overview Installing and using Data Center Fabric Manager Setting up a FICON Director Setting up cascaded FICON Directors FICON Directors in an extended distance environment FICON Directors in an intermixed environment Backing up Director configuration data Backing up DCFM configuration data For information about configuring Cisco MDS FICON Directors. REDP4392. 2005. © Copyright IBM Corp.8 Chapter 8.

8.1. “Configuring a cascaded FICON Director topology” on page 133. In all cases. we show our approach to configuring FICON Directors in different environments. and the implementation of ISLs.2. 8.3. the term switch refers to FICON Director. For switched point-to-point and cascaded FICON Director environments we illustrate how to set up the following: IP addresses (for switch management) Licensed features Logical switches Domain ID and Insistent Domain ID Port-Based Routing (PBR). All examples are based on an IBM System Storage SAN768B Director (2499-384) and an IBM System Storage SAN384B Director (2499-192). Therefore. “Configuring a switched point-to-point topology” on page 115 and Chapter 7.3 and on Fabric OS (FOS) Version 6. The only differences are the link addressing scheme on the System z server side.0e.1. Configuration for all other B-type products is similar. the need for the high-integrity fabric function. 8. “Setting up a FICON Director” on page 164 applies to both topologies. Terminology: Throughout this chapter the terms FICON Director and switch are used interchangeably.1 Configuration flowchart The flowchart displayed in Figure 8-1 on page 153 illustrates the approach we used to configure our FICON Directors for the environments discussed in Chapter 6.8.4. “Planning the FICON environment” on page 59. The window captures are based on Data Center Fabric Manager (DCFM) Version 10. “Setting up cascaded FICON Directors” on page 190 only applies to cascaded FICON Directors. Configuring switched point-to-point and cascaded FICON Directors topologies are very similar. Dynamic Load Sharing (DLS) and In-Order Delivery (IOD) Control Unit Port (CUP) Port type and speed Buffer credits Allow/Prohibit matrix Zoning Port fencing Inter-Switch Links (ISLs) including Traffic Isolation (TI) and Quality of Service (QoS) High-integrity fabric and policies (fabric binding) The configuration described in this chapter is based on considerations explained in Chapter 4.1 Configuration overview In this chapter. 152 FICON Planning and Implementation Guide .

1. we installed the Data Center Fabric Manager (DCFM). “Configuring Port Fencing (optional)” on page 188 8.3.3.8. you must configure the Switch Connection Control (SCC) as described in 8. “Setting up Inter-Switch Links” on page 190 8.3.4. IOD. DCFM was used to set up our FICON Directors.2. “Setting up a high integrity fabric” on page 192. For controlling and maintaining the IBM 2499 FICON Directors.1. “Setting up zoning (optional)” on page 185 8.3. you would only have one WWNN entry (from that particular FICON Director). “Setting up the Allow/Prohibit Matrix (optional)” on page 184 8.4. “Changing buffer credits” on page 183 8. “Configuring the Domain ID and Insistent Domain ID” on 8. and the last two steps are only relevant to a cascaded FICON Directors topology.3. some optional steps may not be needed. if you plan on installing a single FICON Director and want to use the 2-byte link addressing on the System z server for future use in a two-site configuration. “Enabling the Control Unit Port” on page 177 8. “Setting up PBR. Before we started with the configuration of the FICON Directors. Configuring FICON Directors 153 .5.7. “Setting up a high integrity fabric” on page 192 Figure 8-1 Configuration flowchart Depending on your requirements. as well as to configure some optional functions. both CP cards in each chassis must have connectivity to the DCFM server. and DLS” on page 175 8.10.3.11. In the SCC policy. a total of three Chapter 8.3.2. “Changing the IP addresses” on page 164 8. “Setting up a logical switch (optional)” on page 170 8.8.3.3. For each FICON Director.3.6. “Changing port type and speed” on page 179 8. “Enabling features (optional)” on page 169 8.2.3. 8.4. However.3.4.2 FICON Director management connectivity The FICON Directors and DCFM servers should always be interconnected on a separate LAN to isolate Director management traffic from other IP traffic.1.9.

62 10. Note: IP addresses in the range of 10.1 Firewall Corporate Network System Storage SAN384B Figure 8-2 Our FICON Director management environment In this configuration.77. and one for the Director.1.1. the DCFM server and all CP cards in the FICON Directors are connected through an Ethernet LAN (known as the service LAN).1.1.1.255 are reserved for the internal communication of the IBM 2499 FICON Directors and must not be used for the CP cards or Directors. 8.IP addresses are required: two for the CP cards.1. The first three octets of the subnet mask for the service LAN are set to a value of 255. In our case.0.0. and firewall/router.x to ensure that it is isolated from the corporate network.77.1.10 CP0 CP1 SW 66 10. we are using 255. Our FICON Director management environment is shown in Figure 8-2. Because the DCFM server and client applications do direct polling of the fabric to gather configuration and status information. It also shows the IP addresses used for our FICON Directors. which resided in the corporate network.1.0.32 10.255. One IP address is required for each DCFM server.1.2 Installing and using Data Center Fabric Manager The Data Center Fabric Manager (DCFM) software is used to manage several Directors and switches through one graphical user interface (GUI).255.20 SW 65 CP0 CP1 10.61 10. Directors. A router with firewall capabilities was used to allow only remote access to the FICON Directors and the DCFM server from the DCFM client.255.1. 154 FICON Planning and Implementation Guide . and switches that are installed and connected via the network to the Management Server.0.77.1.30 10. Any Director the DCFM server and client cannot reach will go grey within their respective DCFM application. DCFM server. With the DCFM you are able to configure and monitor all fabrics.1.1 to 10.255.21 10.0.1.1. The forth octet (x) of the subnet mask can be adjusted based on the number of FICON Directors and DCFM servers that are connected to the service LAN. Data Center Fabric Manager Server/Client Data Center Fabric Manager Client System Storage SAN768B 10.1. they must be able to access every Director in the fabric.31 10.77.1. which will support up to 254 IP addresses.22 10.

Figure 8-3 Installation Complete window Ensure that Launch DFCM Configuration is selected. select the appropriate field and then type the location of the installation you wish to import.The DCFM is a client-server management software package. a Fabric Manager (FM). the installation will start. Configuring FICON Directors 155 . double-click the installation package from the installation source (located in folder Windows\install. all other users can download the client part of the package by typing the server’s IP address in a Web browser address field.ibm.1 Installing Data Center Fabric Manager Follow these steps to install DCFM on the server: 1.2. If you do not need to migrate data from a previous installation. After the server software is installed. Chapter 8. click Next. then click Done. then select No and click Next (see Figure 8-4 on page 156). 2. Get the installation DVD or the installation package for DCFM.You can import data from an Enterprise Fabric Connectivity Manager (EFCM). Otherwise. If autorun is enabled. On the Welcome window. 3. Follow the instructions shown on the window until you reach the Installation Complete window shown in Figure 8-3.com/common/ssi/pm/sp/n/tsd03064usen/TSD03064USEN.PDF 8. Check the Data Center Fabric Manager (DCFM) data sheet for hardware and software prerequisites at: ftp://ftp. or from a previous Data Center Fabric Manager (DCFM) installation. 4.software.exe). If you are migrating data.

Select Internal FTP Server. then your installation does not require a license.Figure 8-4 Data migration window 5. and click Next. The License Key field is not case sensitive. Enter the serial number (from the DVD box) and the Server License (from the Key Certificate). If the window does not appear. Figure 8-5 Select FTP Server 156 FICON Planning and Implementation Guide . as shown in Figure 8-5. 6.

Also. make sure that it is configured in the DNS server.1. because if it is selected. 8. do not use port 2638. select Software Configuration → FTP/SCP and make the required changes. Note: If a firewall is installed between the DCFM server or client and the FICON Directors. Keep in mind that 16 consecutive ports will be needed from the defined Starting Port #. The Syslog Port # must remain as 514. because this port number cannot be changed in the FICON Director. In the Options window. Enter the appropriate port numbers (see Figure 8-7 on page 158).0. then the configured ports need to be defined in the firewall to allow that traffic flow. From the pull-down menu in Figure 8-6. Important: Do not select 127.10). clients will not be able to connect to the server’s Web application to download the client part of the software. select the IP address that will be used by the clients as Return Address (10.1. click Next. After making your selection. If only one network adapter is installed in the server. Chapter 8.1 as Return Address. Configuring FICON Directors 157 . if you are not using the default values.10) and the address that will be used to connect to the Directors and switches as Preferred Address (10. If you would prefer to select the server name.0. then the address will be the same for both pull-down menus.1. Figure 8-6 Select communication interface 9. because it is used internally by the DCFM server.1. 7. as shown in Figure 8-6.Note: You can change the FTP Server settings later in the DCFM by selecting SAN → Options.

which will be managed by the DCFM server as shown in Figure 8-8 on page 159. To change the SNMP Port #. In the Options window.Figure 8-7 Configure network ports Tip: You can change these port numbers later in the DCFM by selecting SAN → Options. use Monitor → SNMP Traps. 10.Click Next and select the size of your FICON/SAN installation. select Software Configuration → Server Port to make the changes. 158 FICON Planning and Implementation Guide .

12. then the user is administrator and the password is password. If it is a new install.Wait until the database is initialized. then click Next at the Server License Summary window.Click Next after you have made the selection.Figure 8-8 Select SAN size 11. If this is a migration or upgrade from a previous management software installation. Configuring FICON Directors 159 .Wait until the login window appears (see Figure 8-9). 14.At the Start Server window select Start DCFM Client. Click Login. then you need to use the user and password from that version. then click Finish. 13. Figure 8-9 DCFM Login window Chapter 8.

Refer to 8. Figure 8-10 DCFM first startup 16.2. Follow the instructions shown on the windows until Figure 8-11 on page 161 is displayed. 3.2. In the Options window. Double-click the installation package from the installation source (located in folder odbc\Windows\install. Figure 8-10 will appear. 160 FICON Planning and Implementation Guide .You will see the DCFM Login Banner. The ODBC Driver needs to be installed on every personal computer that will generate customized reports. select Security Misc. Installing the Open DataBase Connectivity (ODBC) Driver (optional) Now you can install the ODBC Driver.exe).Select SAN → Options. This can be done after clicking OK. to Microsoft® Excel). Follow these steps to install the ODBC Driver: 1. “Using the DCFM” on page 161 for an explanation of how to change the default password. which you can change or remove if it is not needed. 2. and change the text in the Banner Message field or just remove the check mark. if it is not needed. Get the installation DVD or the installation package for DCFM.15. to be able to export data from the DCFM database to generate customized reports (for example.

Configuring FICON Directors 161 . Click Next. 8. At the Choose Data Source panel.2. 4. Enter the server IP address from the DCFM installation (in our example. you can always use the help button in the DCFM.1. To import data from the DCFM database and create a customized report in Microsoft Excel®. Select a sort order.10). Select the cell you where want to enter the queried data and click OK.2 Using the DCFM In this section we explain functions in the DCFM that are needed during setup. The installation is complete.) and click OK. Open Microsoft Excel and select Data → Import External Data → New Database Query. 2. 3. For other functions. follow these steps: 1.Figure 8-11 ODBC configuration window 4. Click Next when done. 5. 10. If you need to refresh the queried data of a saved form. The Port Number is 2638 (this port is automatically assigned during DCFM installation and cannot be changed). Click Finish.1. if needed. Click Next when done. The last panel will give you the possibility to save your query setup. Chapter 8. Now you can select the required entries and add them to your selection by clicking the arrow pointing to the right. In the Filter Data panel you can set filters. select Data → Refresh Data. if needed. then Done. Click Next after your selections are made. 7. 6. select DCFM* (this entry was created with the installation of the ODBC Driver. 5. The user is guest and the default password is password.

If you want to change the admin password. you can build a new group that includes only the Directors that you assign to it. Click Add to add a Director. username (admin). 4. as shown in Example 8-1 on page 163. Click Edit at the Users field if you want to change the password. You can also download the required Java Version and the MIB tables if needed for external system monitoring software. Accept the default values. go to the DCFM program directory in the command prompt and enter the following command dbpassword <Username> <Password> <New_Password> <Confirm_Password>. Click Web Start the DCFM Client. Click the Change Password button to change the password for the selected user.Discovering a FICON Director Before you can manage a FICON Director.) Using the DCFM client To use the client from any personal computer. During the configuration. so that you do not need to download the client again. You can also create new user roles or edit an existing user role at the Roles field. select Edit and change the password for the selected Director. 1. You can also provide a Fabric Name. The Director will show up in the lower part of the Discover Setup window. Go to Discover → Setup in the menu bar of the DCFM. You will reach the welcome window from the DCFM server. If you click Add in the Resource Groups field. To change the passwords on the Director. Creating and changing users and passwords To change the password. Select the SNMP tab. (Note that this password change is only for the communication between DCFM and the Director. Enter the Director’s IP address. 2. The handling is exactly the same as at the server. click Add and assign the user to a Resource Group by clicking the arrow pointing to the right. then click OK after the password is changed. A shortcut will be created in your program list. you must open a Web browser and enter the IP address of the DCFM server in the Web browser’s address field. To change the ODBC password on the DCFM server. it may happen that you will have to rediscover a Director. Click OK after the password is changed. Select the Director and click Discover to rediscover it. A new window appears. then click OK. and the password (the default is password). or create a new user for the DCFM. go to SAN → Users at the DCFM menu bar. 5. To create a new user. Click Close to leave the discovery dialog. Click OK to activate all your changes. it must be discovered. 162 FICON Planning and Implementation Guide . right-click the Director in the DCFM and select Element Manager → Admin. Click OK after the password is changed. and select v3 in the SNMP Version drop-down menu. click the User tab. select the discovered Director and click Delete. In the Switch Administrator window. To do this. The FICON Director is now discovered and can be managed via DCFM. The client will be downloaded to your workstation and then the login window will appear. Enter the new Fabric Name and click OK. Click Apply to activate the changes. The password itself will be changed via the Element Manager of the Director as described in “Creating and changing users and passwords”. 3.

right-click the Fabric and select Properties.1. Configuring FICON Directors 163 . To change the Fabric Name. You can change the Name. You can also change the port names by clicking the Port tab in the properties window and clicking in the Name field to change the port name. Renaming the Director and ports To change the Director name. Select Display and click the Set Up FICON Display button. Select the Port in the Port Administration window and click the Rename button. To change the Port view. by right-clicking it and selecting Element Manager → Hardware. You can also display the node descriptor for a Director.Example 8-1 Changing the ODBC password C:\Program Files\DCFM 10. The Name Server window opens and you will see all connected devices. right-click the Director in the DCFM and select Properties. click SAN → Options. The port name can also be changed by right-clicking the Director and selecting Element Manager → Ports. Double-click in the Name field and type the new name. Changing the view for FICON management To change the view for FICON. select the hex display option from the pull-down menu on the toolbar as illustrated in Figure 8-12. Figure 8-12 Set to hex Viewing node descriptors If you have changed the view for FICON management. Type the name for the port and click Rename to save the changes. Setting the display to hex To view the display in hexadecimal. Click OK after the changes are done. Click OK after all changes are done. you will see the node descriptor details in the DCFM for each port on the left side of the window. Description. The Option window appears. By selecting the list entry and clicking Detail View. Click OK to close the window. Chapter 8. right-click a Director (on the left of the window) and select Port Display. as explained in “Installing the Open DataBase Connectivity (ODBC) Driver (optional)” on page 160. Location and Contact by double-clicking the desired field. click Name Server on the left side of the window. The properties window opens. In the Directors Element Manager. Check-mark the details you want to display.3\bin> Now you need to uninstall the ODBC driver on all workstations that use this function and install it again with the new password.1.3\bin>dbpassword guest password test1test test1test DB is updated Successfully C:\Program Files\DCFM 10. you will get a detailed view for the selected device.

Connect a personal computer or laptop to the active Control Processor Card (CP0 in Slot 6 or CP1 in Slot 7). then the serial port will have to be used. Service. Right-click the Network Adapter and click Properties. which applies to Directors in both topologies. Additional configuration steps for a cascaded FICON Director are covered in 8. Steps to configure the Directors IP addresses using the serial port are described in IBM System Storage SAN768B Installation. Follow these steps to change the IP addresses in the Director: 1. The steps to configure the Directors IP addresses are as follows: Connecting a laptop or personal computer to the Director’s active Control Processor (CP) Configuring the laptop or personal computer to be in the same network Altering the Director’s IP addresses Connecting the Director to the customer network Note: If you are redeploying a FICON Director and do not know its IP addresses.1 Changing the IP addresses Before connecting the FICON Director to the network. 3.8. GA32-0574 and in IBM System Storage SAN384B Installation.4. 2. 8. you need to change the default IP address to match your network layout and rules. We also include optional steps for the most commonly used functions. The active CP is indicated by the blue LED on the card. GA52-1333. Use the Management Port to connect the network cable. as shown in Figure 8-13 on page 165.3 Setting up a FICON Director In this section we illustrate how to set up FICON Directors. Alter the personal computer or laptop IP address to be in the same network as the Director.3. and User's Guide. Go to Start → Control Panel → Network Connections (use the classic view in the control panel). and User's Guide. 164 FICON Planning and Implementation Guide . “Setting up cascaded FICON Directors” on page 190. Service.

Configuring FICON Directors 165 . Figure 8-14 Select Protocol 5. in this case) to be in the same network as the Director.77.70. Click OK on both Properties windows. Select Internet Protocol (TCP/IP) and click Properties (see Figure 8-14).77. Select Use the following IP address and enter the address (10. 6.Figure 8-13 Select network interface 4. see Figure 8-15 on page 166. Chapter 8.

6 or later). In the login window. enter the default user (admin) and the default password (password).77. select Launch and click Edit. The default IP address for CP0 in Slot 6 is 10. then click OK. you will need to install a new Java version (Version 1.77. in this case) and press Enter. Click Advanced. Enter the default IP address in the Web browser’s address field (10.77.74.Figure 8-15 Setting up a laptop IP address 7. The default IP address for CP1 in Slot 7 is 10. go to Start → Control Panel → Folder Options (in classic view). proceed to step 9 on page 167. If you do not receive this message.6 is installed.75.77. Connect the cable to the active CP Card in the Director and open a Web browser. The window shown in Figure 8-17 on page 167 will appear.77.77. 166 FICON Planning and Implementation Guide . In the Edit File Type window. (The Mozilla Firefox Web browser will ask for open or save.) If you receive a message like the one shown in Figure 8-16.77. Figure 8-16 Wrong Java version message If Version 1. then scroll down to and select JNLP. Select the File Types tab. 8. Select open with and click OK.

0_05\bin\javaws. The login window should appear after you enter the IP address in the Web browser’s address field. Select the Network tab at the Switch Administration window to reach Figure 8-18 on page 168. Figure 8-17 Assign Java version to JNLP files Click OK after the changes are made.In the field “Application used to perform action” you must change the location of your Java version. for example.6. 9. Configuring FICON Directors 167 . At the Director’s graphical user interface. Type in the user (admin) and password (password) and click OK. C:\Program Files\Java\jre1. Chapter 8. select Manage → Switch Admin.exe" "%1.

1. Our example is based on Figure 8-2 on page 154. 11.22 255. Table 8-1 IP address example Virtual address Ethernet IP Ethernet mask Gateway IP 10.1.20 255.255 cannot be used because they are already used internally in the Director. as described in “Discovering a FICON Director” on page 162. IP addresses between 10.1.0.1.Connect each Control Processor Management Port to the network (Ethernet switch). When connecting with the IP address of CP0 or CP1.Figure 8-18 Setup management IP address 10.255.255.0 n/a CP1 address 10.0.0.1.255. The Director needs to be discovered by DCFM.255.0 10.255.To activate the new IP addresses.Fill in the required fields.1. This will guarantee a connection even if a Control Processor switchover has occurred (for example.21 255. 168 FICON Planning and Implementation Guide . only the active CP is manageable. during a firmware upgrade). scroll down to the bottom of the window and click Apply.0 n/a Important: The Virtual address is later used to connect to the Director. 12.0 to 10.0.255. Table 8-1 lists the IP addressing scheme we used.1.1 CP0 address 10.1.

then enter the license key for the feature you want to enable. Configuring FICON Directors 169 . Chapter 8.2 Enabling features (optional) To use a licensed feature. Note: The license key is case sensitive and will only work on the Director to which it is assigned. 8. as shown in Figure 8-20 on page 170. you need to compare the WWN on the license with your Director. 1. Go to the Switch Admin window in the Element Manager of the Director. Click Add. Figure 8-19 License overview 3. refer to the remaining parts of this section. as shown in Figure 8-19. it must first be enabled. Select the License tab. This can be done via DCFM by right-clicking the Director and selecting Configure → Element Manager → Admin. 2.Using the FICON wizard The FICON wizard provides you with a simple means to set up the following configuration options: Insistent Domain ID Set high integrity fabric Enable port-based routing Enable in order delivery Disable dynamic load sharing Enable the FICON Management Server (FMS) for Control Unit Port (CUP) management You can access the FICON wizard via DCFM by right-clicking the Director and selecting Configure → FICON → Configure Fabric. Therefore.3. If you decide not to use the FICON wizard or you need to configure other options not included in the FICON wizard.

which is disruptive. 170 FICON Planning and Implementation Guide . Important: If the Firmware was updated to level 6. then you have by default one logical switch. select the Director in DCFM that you want to enable for Virtual Fabric in the Chassis Group. then you need to enable Virtual Fabric first. 2. If you have more than one license. If the Director uses 48 port cards. Click Add License to activate the license. If the Director was delivered with Firmware level 6.3 Setting up a logical switch (optional) A logical switch in the Director can be used to separate the physical ports for different System z environments or for mixed environments (FICON and Fibre Channel). then repeat step 3 on page 169 for each. Two of these switches can have FICON Management Server and CUP enabled. The default logical switch has Logical Fabric ID (FID) 128 assigned. as shown in Figure 8-21 on page 171. A maximum of eight logical switches can be created. you must create two logical switches with Zero-based Area Assignment. Right-click the Director and select Enable Virtual Fabric. but it does not need to have ports assigned to it. If Virtual Fabric is enabled (indicated by the blue V on the Director in the DCFM Switch view). thereby allowing you to create more.2.2.Figure 8-20 Add a license 4. Follow these steps to create a logical switch: 1. The feature is now enabled and can be used or configured. The default logical switch is always present if Virtual Fabric is enabled. then you can configure a logical switch and assign all ports from the default logical switch to the new logical switch. 8. If you want to assign port IDs (PIDs).0. Otherwise.0 (required for logical switches). then proceed to step 5 on page 171.3.

Configuring FICON Directors 171 . which is disruptive to the Director. 5. Select Configure → Logical Switches at the DCFM menu bar. because the Director was changed and appears as a new Director. You need to discover the Director again after the reboot. 4. Refer to “Discovering a FICON Director” on page 162 for more details about this topic. 6. In the Chassis drop-down menu. Chapter 8.Figure 8-21 Enable Virtual Fabric 3. Read the warning message and select OK. Important: Clicking OK will immediately reboot the Director and create a logical switch. 7. select the chassis where you want to create a new logical switch. Wait for the Director to complete the reboot. as shown in Figure 8-22 on page 172. Wait for the Logical Switches configuration window to appear.

Figure 8-22 Create a Logical Switch 8. The FID must be unique in the selected chassis. but only in Brocade Native Interoperability Mode or Port Based 172 FICON Planning and Implementation Guide . then click the New Switch button. The 256 Area Limit can be set to Zero-Based Area Assignment. Select Undiscovered Logical Switch. see Figure 8-23. Enter the desired settings for the fabric using the Fabric tab. because this is not supported by FICON. Note: Keep in mind that only logical switches with the same FID can be cascaded. Figure 8-23 Logical switch configuration The Logical Fabric ID (FID) can be a number from 1 to 128 (decimal). Ports will be numbered in ascending order. starting from 00 to FF. 9. Do not select Base Switch or Base Fabric for Transport. Ports with a Port Index above 255 can be added. and do not set the 256 Area Limit to Disable. for each logical switch.

11. Configuring FICON Directors 173 . then click Start. Chapter 8.Click OK after all settings are complete. The default Domain ID is 1. the Director must be disabled. 10.Area Assignment (ports with a Port Index above 255 cannot be added. After the status Success displays. as explained here. 12. The logical switch is now created and can be used like any other physical switch or Director. and click the arrow pointing to the right. and select Re-Enable Ports after moving them (if not selected.Select the Switch tab and assign the Domain ID and the logical switch name. 15. click Close.3. which is decimal 1 to 239. 8. see Figure 8-24. 14.Now you need to assign the desired ports to the new logical switch. instead. Only Domain IDs between 1 and EF in hex are possible. Instructions that explain this process are provided in “Discovering a FICON Director” on page 162. 16. Select the ports on the left side.Wait for the changes to complete. click OK. the default Port numbering will be used). This will allow you to later merge two Directors without disruption.Verify the changes in the confirmation window. Figure 8-24 Assign ports to a logical switch 13. Also.4 Configuring the Domain ID and Insistent Domain ID To set up the Domain ID. all ports must be enabled manually). You can change the port numbering for logical switches (except for the default logical switch) later in the Port Admin panel in the Element Manager by using the Bind PID and Un-Bind PID buttons. We also explain how to set the Insistent Domain ID flag.At this point you need to discover the new logical switch. check-mark the Insistent Domain ID field.After the ports are assigned.

Follow these steps to change the Domain ID to the desired value: 1. Click the Configure tab. 9. Go to the Switch Admin window in the Element Manager of the Director. Now you can change the Domain ID field to the desired value. 4. which is 101 in decimal. 7. Select Disable in the Switch Status field and click Apply. Click Apply. The value has to be entered in decimal. but the Domain ID in the Director is defined with a decimal value. 174 FICON Planning and Implementation Guide . click Yes to confirm the action. 6. as shown in Figure 8-26 on page 175. Be aware that the switch address is configured in hex. Note: The Domain ID must match the switch address configured in HCD/IOCP on the System z side. Click Insistent Domain ID Mode to check-mark it. After reading the warning message. Read the warning message. In our example it will be 65 in hex. 2. as shown in Figure 8-25. 5. Figure 8-25 Domain ID configuration 3. and confirm it by clicking Yes. If not already done. 8. You can also change the Name of the Director so it can be identified easily. click Show Advanced Mode in the upper right corner of the Switch Administration window. Select the Switch tab. Important: Setting the Director to disable is disruptive. This can be done via the DCFM by right-clicking the Director and selecting Configure → Element Manager → Admin.

Chapter 8.Select Enable in the Switch Status field and click Apply. it must be set to disabled. If the Director is already disabled. Instead. Follow these steps to set up PBR. This can be done via the DCFM by right-clicking the Director and selecting Configure → Element Manager → Admin.) 12. Select Disable in the Switch Status field and click Apply. (You may also skip this step and configure Port-Based-Routing (PBR). the new Director will be segmented until another Domain ID is set at the joining Director. Dynamic Load Sharing (DLS) and In-Order Delivery (IOD). Return to the Switch Admin window in the Element Manager of the Director. IOD.Click Apply to save the changes.3.Read the warning message and click Yes to confirm the action. you must disable creditrecovery at all ISLs. Dynamic Load Sharing (DLS) can be used in a FICON environment if lossless DLS is also enabled. IOD. otherwise. 2. and DLS: 1. 8. which will be the case if an ISL is added or removed.Figure 8-26 Configure Insistent Domain ID 10.5 Setting up PBR. go to step 6 on page 176. Configuring FICON Directors 175 . Lossless DLS will ensure that no frames are dropped during rebalancing of the Inter-Switch Links (ISLs).To enable the Director at the Switch Status field. 11. Otherwise. before you enable the Director. Note that it cannot be automatically changed if another Director joins the Fabric with the same Domain ID. 3. return to the Switch tab. and DLS Port-Based Routing (PBR) and In-Order Delivery (IOD) must be set in a FICON environment. When lossless DLS is used in long distance environments. lossless DLS is not supported in long distance environments. 13. The Domain ID is now set. However. select the Switch tab.

In the DCFM server or client. Select the Routing tab. Enter. Figure 8-27 Set up Port-Based Routing. If you have set DLS to On. admin) and press Enter. then you need to enable lossless DLS also via the command-line interface (CLI). Type logout and close the window.Important: Setting the Director to disable is disruptive. if needed. d. 5. setcontext 1 to connect to a logical switch with FID 1. 176 FICON Planning and Implementation Guide . right-click the Director and select Telnet. g. For long distance environments. The following steps explain one method of using the CLI: a. To enable the Director at the Switch Status field. return to the Switch tab. Set In-Order Delivery (IOD) to On. DLS and IOD 7. proceed to step 10. Click Show Advanced Mode in the upper right corner. Click Apply to save the changes. 4. f. e. Type the command iodset --enable -losslessdls and press Enter To check the settings. and confirm it by clicking Yes. A new window opens. c. (You may also skip this step and configure the Control Unit Port (CUP) before enabling the Director. enter iodset --show and dlsshow. 6.) 10. Select Port-Based-Routing in the Advanced Performance Tuning (APT) Policy field. for example.) b. Enter the user name (in our case. Set Dynamic Load Sharing (DLS) to Off if lossless DLS should not be used. Read the warning message. as shown in Figure 8-27. 8. Enter the password and press Enter (the default is password).Select Enable in the Switch Status field and click Apply. Otherwise. execute portcfgcreditrecovery --disable slot/port for each ISL. (You can also use an ssh connection to the Director. 9.

In this case. as shown in Figure 8-28 on page 178. Click Show Advanced Mode in the upper right corner. 3. Select Disable in the Switch Status field and click Apply. 2. The Director must be enabled before you can disable the ports. Go to the Switch Admin window in the Element Manager of the Director.7.3. ports FE and FF in a 256 Port Director must be disabled because those port addresses are used by the CUP server. Refer to 8. The CUP can be enabled while the Director is enabled. Follow these steps to enable the CUP: 1. because our example is in the process of setting up the Director. Select the FICON CUP tab.6 Enabling the Control Unit Port If the Director has a CUP license. and confirm it by clicking Yes.Read the warning message and click Yes to confirm the action.11. you can do it while the Director is disabled. 6. 5. then the Control Unit Port (CUP) can be enabled.3. Chapter 8. If the Director is already disabled. The Director is now ready to be used in a FICON environment. This can be done via the DCFM by right-clicking the Director and selecting Configure → Element Manager → Admin. Read the warning message. “Changing port type and speed” on page 179 for more information. proceed to step 6. 4. A Registered State Change Notification (RSCN) will be sent to inform the System z server that the CUP is now enabled. Otherwise. Configuring FICON Directors 177 . Important: Setting the Director to disable is disruptive. 8. select the Switch tab. However.

Figure 8-28 Enable the CUP 7.Read the warning message and click Yes to confirm the action. The CUP Port is now usable from the System z server. or to change it.Select Enable in the Switch Status field and click Apply. A new window opens. Enter the password and press Enter (the default is password).Click Apply after the changes are made. 178 FICON Planning and Implementation Guide . Click Apply.Return to the Switch tab.Now you can alter the FICON Management Server Behavior Control (Mode Register) field. 10. Select Enable in the FICON Management Server Mode field. 12. for example. Enter. you need to connect via telnet or ssh to the Director’s IP address. right-click the Director and select Telnet.) 2. (You can also use an ssh connection to the Director. 14. Missing Interrupt Handler Process Timeout (MIHPTO) The default timeout value for the CUP is 180 seconds.Read the message and click Yes to confirm the action. if needed. setcontext 1 if you need to connect to a logical switch with FID 1. 3. To check the value. 15. 8. to enable the Director at the Switch Status field. 4. 13. Here is one method for accomplishing this task: 1. Select the desired functions you want to use with the CUP. admin). 9. In the DCFM server or client. Read the message and click Yes to confirm the action. see Example 8-2 on page 179. Enter the user name (in this case. then press Enter. 11.

Important: Changing the port configuration is disruptive to port operations. The MIHPTO for the CUP is now set and active. 2. 3. MIHPTO has been changed to 180 seconds IBM_SAN768B:FID128:admin>logout 5. Click Show Advanced Mode in the upper right corner to see the configuration options.Example 8-2 FICON CUP MIHPTO IBM_SAN768B login: admin Password: ----------------------------------------------------------------IBM_SAN768B:FID128:admin> ficoncupshow mihpto MIHPTO for the CUP: 180 seconds IBM_SAN768B:FID128:admin> ficoncupset mihpto 181 The input value of the MIHPTO will be rounded down to the nearest value divisible by 10. Chapter 8. E_Port. 8. Click Edit Configuration for the selected port. Enter logout and close the window. 4. Configuring FICON Directors 179 . or L_Port device is connected. 6. Follow these steps to configure a port: 1. A value between 15 and 600 seconds can be used. see Figure 8-29 on page 180. The Port Administration window will appear. you must modify the port configuration. Go to the port configuration window of the Director’s Element Manager. Right-click the Director in the DCFM and select Element Manager → Ports. Important: This is disruptive for the traffic on the selected port. which automatically detects if an F_Port. or will be rounded down. This value must be divisible by 10 if over 63. This is due to IBM's MIHPTO specification. enter ficoncupset mihpto xxx. where xxx is the new value. or Extended Distance Link.7 Changing port type and speed To change a specific port type. The default speed is set to auto-negotiate. Read the information at the confirmation panel and click Yes.3. enter ficoncupshow mihpto. speed. To see the timeout that is currently set. To change the value. The default port type is a U_Port. 7. Select the port you want to configure and click Disable. 5.

F_Port is for all other devices. 7. which is not supported for a FICON environment. Be aware that EX_Port is not selectable. or accept the defaults if you only want to change the speed. Select the Port Type you want to configure. 180 FICON Planning and Implementation Guide .E_Port is for ISLs. because it is only available at the Base Switch in a logical switch configuration. .Figure 8-29 Port type configuration 6. . Note the following points: .L_Port is for loop devices. Click Next and Figure 8-30 on page 181 will appear.

Chapter 8. which is only valid for 8 Gbps SFPs. Configuring FICON Directors 181 . to limit the throughput of the port (200. For more detailed information about this topic. 8. 600. you can configure the distance of the link (L0:normal is the default value). If 4 Gbps SFPs are present.The Ingress Speed Limit (MBps) drop-down menu is a Quality of Service function. according to the specifications. 9. “Changing buffer credits” on page 183. 4000. These interfaces use idles to fill the link. refer to 8.8. 1000. 4.With Long Distance Mode.2) is set to ARB(FF). At the Confirmation window.At the Speed drop-down menu. 3500. idles will be sent even if ARB(FF) is configured. 400. you can select Auto (for auto-negotiate) or 1. Verifying and changing the fill word A fill word is sent when there is no traffic on the link. For 8 Gbps SFPs. and 4 Gbps. click Save after verifying the settings. the fill word should be ARB(FF). it is useful to give the port a name by clicking Rename. 2 Gbps. . 10. 2000. 800. .Figure 8-30 Port speed and distance configuration Note the following points: . 2. and 8 Gbps. 5000. However. The default (when a Director is delivered with FOS 6. some interfaces were developed before the final specifications were approved. 11. Click Next after all selections are done.3. The mode can only be verified and changed via the CLI. After the port type and speed are configured.Click Close after the configuration complete message. which is the standard for 1 Gbps. 1500. in order to keep it online.Click Enable to enable the port after the configuration is complete. 7000 and 8000 Mbps are possible).

right-click the Director and select Telnet. Port8 to Idles. Enter. for example. 182 FICON Planning and Implementation Guide . it will look as shown in Example 8-5.i<32.Important: Changing the mode is disruptive for the port. Enter the password and press Enter (the default is password). as shown in Example 8-4.i++)). you must replace 16 with 32.2. Port8 (see Example 8-3). 4. Example 8-5 Script to change the fill word for a 32-port card in slot 3 for ((i=0. 6. You can also use a script that changes all ports on a given port card. and the Director will use the configured mode. The fill word is now changed. do (echo 3/$i. Enter logout and close the window.0e_rc1_bld01 DCXBOT login: admin Password: ----------------------------------------------------------------IBM_SAN768B:FID128:admin> portcfgfillword 1/8 0 IBM_SAN768B:FID128:admin> portcfgshow Ports of Slot 1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 -----------------+--+--+--+--+----+--+--+--+----+--+--+--+----+--+--+-Speed AN AN AN AN AN AN AN AN AN AN AN AN AN AN AN AN Fill Word 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 ~~~ IBM_SAN768B:FID128:admin>logout 7. The portcfgfillword 1/8 0 command changes the fill word for Slot1. The portcfgshow command will list all ports of the Director. For a 32-port card in slot 3. The portcfgshow 1/8 command will only list the settings for Slot1. done This script will change the settings for the first 16 ports at the card in Slot 2 to ARB(FF).i<16. setcontext 1 if you need to connect to a logical switch with FID 1. (You can also use an ssh connection to the Director. do (echo 2/$i.portcfgfillword 2/$i 1). If a 32-port card is used. The portcfgfillword 1/8 1 command changes the fill word for Slot1. In the DCFM server or client. admin) and press Enter. The slot also needs to be changed in both positions in the script.i++)). Example 8-3 Changing fill words Fabric OS (DCXBOT) Fabos Version 6. if needed. The following method illustrates one way to change the fill word for a 8 Gbps interface: 1. done 8. Example 8-4 Script to change the fill word for a 16-port card in slot 2 for ((i=0.) 2. A new window opens. 5. 3. which is specified by the 1 in the script. Port8 to ARB(FF).portcfgfillword 3/$i 1). Enter the user name (in this case.

5. Enable the port by entering: portenable 1/8. Follow the steps in “Changing port type and speed” on page 179. 3. LD:Auto . admin) and press Enter.8 Changing buffer credits Buffer credits are used as a flow control mechanism to avoid overwriting stored frames received in a port. Enter. Follow these steps to change the buffer credits for an F_Port: 1. The value you enter in the Desired Distance (km) field will determine the number of buffers allocated to the port based on a 2 KB frame size.and enter the link length in the Desired Distance (km) field. A total of 16 buffers will be allocated for a 2 Gbps link (a 1 Gbps link uses the default of 8 buffers). 4.3. use the Element Manager of the Director as follows: 1. if needed. For an explanation of how buffer credits work.8. 7. Enter the user name (in this case. Click Enable to enable the port after the configuration is done. To change the buffer credits for an ISL. A total of 46 buffers will be allocated for a 8 Gbps link. Configuring FICON Directors 183 . refer to “Buffer credits” on page 39. The default is 8 buffer credits per F-Port.and enter the maximum link length in the Desired Distance (km) field. refer to “Buffer credits” on page 39 for more information. 3. 6. Chapter 8. 5. A total of 26 buffers will be allocated for a 4 Gbps link. 2. right-click the Director and select Telnet. Disable the port by entering: portdisable 1/8. At the Confirmation window. Click Next after all selections are done. Open the drop-down menu for the Long Distance Mode. 4. This will show the assigned buffers for each port at the card in Slot 1. as well as the remaining buffers for the port group.) 2. Enter the password and press Enter (the default is password). setcontext 1 if you need to connect to a logical switch with FID 1. 8. click Save after verifying the settings. LS:Static . In the DCFM server or client. The buffer credits will be calculated during link initialization of the E_Port and automatically allocated. Click Close after you receive the configuration complete message. A new window opens. In most cases. Select one of the following options: LE:<=10 km (only use if the link is shorter than 10 km). (You can also use an ssh connection to the Director. for example. To display the assigned buffers use the portbuffershow 1 command. extended distances require that the number of buffer credits be increased at the port. 6. This command will change the buffer credits for a given port: portcfgfportbuffers --enable 1/8 80 This command will change the buffers for port 8 in the card in slot 1 to 80: portcfgfportbuffers --disable 1/8 7. The buffer credits will be set for the given distance and the selected speed.

Figure 8-31 Allow/Prohibit configurations 6. Select the CUP Port Connectivity tab in the lower left corner of the FICON CUP tab. 4. Click Show Advanced Mode in the upper right corner.To create a new matrix.10.To make changes directly. as shown in 8. you must provide a new Name and Description in the pop-up window and then click OK. the window shown in Figure 8-32 on page 185 appears. select the Active Configuration and click Edit.9 Setting up the Allow/Prohibit Matrix (optional) To prohibit traffic between selected ports.3. you can use the Allow/Prohibit Matrix. 3. click New. 5. If you use the default zone.3. This can be done via the DCFM by right-clicking the Director and selecting Configure → Element Manager → Admin. zoning must be active. Go to the Switch Admin window in the Element Manager of the Director. 184 FICON Planning and Implementation Guide . Perform one of the following actions: . 2. then create a new zone with all ports. . click Copy.8. The window shown in Figure 8-31 opens. After you make your selection and click Edit or New. To use the Allow/Prohibit Matrix. Follow these steps to change the Allow/Prohibit Matrix: 1. . Select the FICON CUP tab. “Setting up zoning (optional)” on page 185.To copy an existing configuration.

In the DCFM. Click Yes to confirm the action. you can activate the changes by clicking Activate.The Prohibit button will prohibit all connections to and from the selected port. Figure 8-32 shows that port 03 (hex) is prohibited to connect to ports 00. All other ports are allowed to connect to port 3. you do not need to set up zoning. 7. 8. and 02.10 Setting up zoning (optional) Because FICON connectivity is defined in the System z server based on a port address in the FICON Director. you can modify the matrix by clicking in the fields to allow or prohibit a connection. After making the desired changes. if you want to separate some FICON channels and control units (for a test environment). select Configure → Zoning. 8. Configuring FICON Directors 185 . then click Save (not available for the active matrix).Figure 8-32 Allow/Prohibit Matrix At this point. Note the following points: . to create a new configuration. The window shown in Figure 8-33 on page 186 appears. The default zone in the Director is sufficient. However. or Save As. Chapter 8. If you do not want to activate the changes right away. 01. It is the same function as a port disable. follow these steps: 1.The Block button will block the selected port.3. .

To place the selected ports in the new zone in the middle of the window. 186 FICON Planning and Implementation Guide . you want to create. Test_LPAR). 6. If you want to select a range of ports.) 7. The result is shown in Figure 8-34 on page 187. To create a new Zone. Repeat steps 3 to 7 for all zones. Ensure that Domain Port Index is selected in the Type drop-down menu. 5. use the Shift key. because port-zoning is used in FICON environments. Enter a name for the new Zone (in this case. select the Director for which you want to create Zoning. 4. Select the ports you want to include in the zone you just created at the left side of the window. which includes all other ports of the Director (except ISLs).Figure 8-33 Zoning dialog overview 2. click the New Zone drop-down menu and select New Zone. click the arrow pointing to the right. In the Zoning Scope drop-down menu in the upper left corner of the window. 3. In our example. (You may select more than one port at a time by pressing and holding Ctrl key on the keyboard while selecting the ports with the mouse. we created a Rest_of_the_world zone.

as well as the active Zone Config. 13. Configuring FICON Directors 187 . After all zones are created. Enter a name for the new Zoneconfig (in this case.After activation you will see that all active zones. FICON_Zoneset). Verify that all ports are in the correct Zones. Chapter 8. have green markers attached (see Figure 8-36 on page 188). Click OK after you verify all zones’ this will activate the newly created Zone Config. 12.You will reach the Activate Zone Config window. Figure 8-35 Creating a Zone Config 10. you must put them in a Config. The result is shown in Figure 8-35. 11.Figure 8-34 Creating zones 8. Select the zones in the middle of the window that you want to place in the newly created Zone Config.Now you must activate the new Zone Config. 9. Click the New Config button. then click the arrow pointing to the right.Click Yes when you reach the confirmation window. Select the Zone Config (in this case. FICON Zoneset) and click Activate.

which could lead to degraded performance. To configure Port Fencing. click OK. When Port Fencing is configured. as well as the active zones with all zone members. In the DCFM. 8. follow these steps: 1. A bad link will cause error recovery on the attached device. this will lead to recovery for many devices. The window shown in Figure 8-37 on page 189 appears.Figure 8-36 Zoning activated To verify the active zone.If you get a confirmation window. 188 FICON Planning and Implementation Guide .To leave the Zoning dialog. 15. Zoning is now completed and active in the Director. the port will be set offline automatically to keep the fabric stable. the port will have to be manually set online.3. If the link is an ISL. select Configure → Port Fencing from the menu bar. 14. After the link is checked or repaired. You will see the active Zone Config. click the Active Zone Config tab.11 Configuring Port Fencing (optional) Port Fencing is used to force a port offline if a specified error rate is reached. read the warning and click OK.

3. the port will go offline. Click OK to activate the policy on the Director (or fabric). Configuring FICON Directors 189 . When the threshold is reached for the assigned error type. Two selections are presented: m-EOS (which is for McData Directors) and FOS. (Some policies are only for FOS or m-EOS as indicated at the selection from the drop down menu. we selected FOS for our SAN768B Director. or click Add to create a new one. Select the policy on the left part of the window. In this case. 6.Figure 8-37 Port Fencing dialog 2. the Data Center Fabric Manager (DCFM) User Manual. then clicking the arrow pointing to the right. There are two possible policy types in FOS: the Default and the Custom Policy. We selected the default policy. 7. Chapter 8. 5. and the Fabric OS Command Reference Manual for other functions required for your installation. Assign the policy by highlighting it on the left site. You can also change the policy name in this panel.) Click OK after the settings are done. 4. Select the Fencing Policy from the Violation Type drop-down menu (in this case. If you want to change or verify the policy. The policy is added to the Director as indicated by the green plus (+) sign. You have now assigned the policy to the Director. but not covered in this section. and highlighting the Director on the right side (or the fabric). the Invalid CRCs policy). Refer to Fabric OS (FOS) Administration Guide. click Edit.

3. which will interrupt the traffic on the link. port speed. From the list. The FICON Directors that will be merged have the same setup as any other FICON Director. thereby forcing them to use a specific ISL or trunk exclusively. Click the Enable Trunking button. In a cascaded environment. Trunking (optional) If trunking is enabled for an ISL port. select the port that you want to enable for trunking. 2. Right-click the Director in the DCFM and select Element Manager → Ports. Important: This will directly enable trunking for the selected port. In the Port Administration window. which should be named with TI at the beginning of the zone name. click Show Advanced Mode in the upper right corner. 8 Gbps). it will automatically form a trunk. Both sides of the ISL must have the same setup for them to function properly. Follow these steps to create a TI zone: 1. all links in the trunk (maximum 8) must be in the same port group and must be set up with the same speed (for example. To enable trunking. but ISLs are included. The trunk or ISL cannot be used by any other TI zone. 190 FICON Planning and Implementation Guide . TI zoning is set up like regular zoning. Traffic Isolation (TI) zoning (optional) Using Traffic Isolation (TI) zoning. you can isolate traffic for some devices.1 Setting up Inter-Switch Links Inter-Switch Links (ISLs) are used to connect the two FICON Directors.7. you must first complete the tasks in 8. Therefore. The TI zone will not be included in the Zone Config on the right side of the Zoning window (see Figure 8-38 on page 191). If two of four ports have Quality of Service (QoS) enabled. and Quality of Service must be the same on both sides of the ISL. two trunks will be formed with two ports each (even if they are in the same port group): one trunk with QoS. “Changing port type and speed” on page 179 for more detailed information. 8. Add both sides of the ISL or trunk to the new TI zone.8. but select New TI Zone instead of New Zone. the whole trunk will be assigned to the TI zone.3.4 Setting up cascaded FICON Directors A cascaded FICON Director is one that is connected to another Director via ISLs. “Setting up a FICON Director” on page 164 for both FICON Directors. and one trunk without QoS. the link will go offline for a short moment. If this port is online. This means that the port type (E-Port). 3.4. Trunking. the System z FICON channel has to use the 2-byte addressing scheme. 4. See 8. If an ISL is selected that is a member of a trunk. Each ISL port in the FICON Director is configured as an E_Port. which includes the switch address (first byte) and the port address (second byte). follow these steps: 1. Long Distance Mode. For trunking. All fiber optic links in the trunk must be the same length (maximum of 30m difference). Perform steps 1 to 7 in “Setting up zoning (optional)” on page 185.Trunking will only work in Brocade native Interoperability Mode.

Configuring FICON Directors 191 . Right-click the Director in the DCFM and select Element Manager → Ports. you must use WWN zoning instead of port zoning. select the port that you want to enable for QoS. then the devices in the TI zone will not use another available ISL outside its TI zone. To activate one or more created TI zones in the Directors. 4. Click QoS Enable for the selected port. In the Port Administration window. read the message and select Yes.Figure 8-38 TI zoning setup 2. 5. Important: When Failover is disabled. follow these steps: 1. However. right-click the TI zone and select Configured Enabled. When the configured ISL is available again. it will automatically fall back. Compare the settings at the confirmation window and select OK to activate the zoning changes. You can also select Configured Failover. From the list. click Show Advanced Mode in the upper right corner. If Failover is enabled. you can decrease the usable bandwidth for a device. Click OK after the changes are activated. Quality of Service (QoS) (optional) Using Quality of Service (QoS). click Activate in the Zoning window.7. To activate the TI zone. 2. 6. Chapter 8. 3. To verify that the TI zone is active. To enable QoS.3. 4. This is done with the Ingress Rate Limit (MBps) setup already shown in 8. There is also a possibility to use QoS zones. If a second confirmation window appears. click the Active Zone Config tab. 3. “Changing port type and speed” on page 179. 7. the devices in the TI zone will use another available ISL (which is not in the TI Zone).

To use Quality of Service zones, follow these steps: 1. Complete the steps in “Setting up zoning (optional)” on page 185, but select WWN instead of Domain, Port Index in the Type drop-down menu. 2. Right-click the created zone and select QoS Priority → QoS_High or QoS_Low as shown in Figure 8-39. The default is QoS_Medium, for all zones.

Figure 8-39 QoS priority selection for WWN Zoning

3. The name of the zone will be changed to QoSH (for high priority) or QoSL (for low priority) in front of the given zone name. 4. To activate the QoS zone, click Activate in the zoning window. 5. Compare the settings at the confirmation window and select OK to activate the zoning changes. 6. If a second confirmation window appears, read the message and select Yes. 7. Click OK after the changes are activated. 8. To verify that the zone is active, click the Active Zone Config tab.

8.4.2 Setting up a high integrity fabric
To merge the two Directors, the Switch Connection Control (SCC) Policy must be set up to form a high integrity fabric. Therefore, you need to set the policy to Strict, which will distribute the policy only to FOS 5.2.0 or later Directors, resulting in a stable and secure fabric. In the

192

FICON Planning and Implementation Guide

SCC policy, we show how include the World Wide Node Name (WWNN) of the two Directors in the fabric. To do this, follow these steps: 1. Go to the Switch Admin window in the Element Manager of the Director. This can be done via the DCFM by right-clicking the Director and selecting Configure → Element Manager → Admin. 2. Click Show Advanced Mode in the upper right corner. 3. Select the Security Policies tab. 4. On the left side of the window, click FWCP, which is the Fabric Wide Consistency Policy. 5. Select Strict at the SCC Consistency Behavior drop-down menu. 6. Click Apply. Note: This will change the security settings of the Directors. From now on, each System z server needs to use the 2-byte addressing scheme to initialize the link. 7. Select ACL on the left side of the window. 8. Click Edit to set up the SCC Policy. 9. Select only SCC and click Next. 10.Click Modify to add the WWNNs of the two Directors. 11.Select the Director on the left side, and click Add Switch (see Figure 8-40). This will add the WWNN of the Director you are connected to.

Figure 8-40 Creating SCC policy

Chapter 8. Configuring FICON Directors

193

12.In the Add Other Switch field, enter the WWNN of the second Director and click Add. (The WWNN of the Director is shown on top of the Switch Admin window). 13.Click OK, then click Next. 14.Verify the settings you just made, and click Finish. 15.Back in the Switch Administration window, you now see Defined Policy Set. Click Activate to activate the Policy (see Figure 8-41).

Figure 8-41 SCC policy activated

16.Perform steps 1 on page 193 through 6 on page 193 on the second FICON Director. (Set the Fabric Wide Consistency Policy for SCC to Strict.) 17.Connect the ISLs. The ISLs will be initialized, and a cascaded fabric will be created. Remember to set up the ISLs on both sides to the same settings as described in “Setting up Inter-Switch Links” on page 190. The SCC Policy will be distributed to the second Director automatically. In DCFM, you will now see that the two Directors are merged and connected via ISLs, as shown in Figure 8-42 on page 195.

194

FICON Planning and Implementation Guide

Figure 8-42 FICON Cascaded Director DCFM overview

To verify the ISLs, right-click them in the DCFM Directors View and select Properties. You will see all active ISLs and their settings on both Directors, as shown in Figure 8-43.

Figure 8-43 ISL properties window

Chapter 8. Configuring FICON Directors

195

If ISLs are missing, go to the Port Admin window of the two Directors and check the settings on both sides. Use the Advanced View to see all settings. Scroll to the right side and check the Additional Port Info field (this field also displays some error reasons for the ISL).

8.5 FICON Directors in an extended distance environment
In an extended distance environment, using a Wave Division Multiplexing (WDM) platform will require you to set up the Director to match the WDM configuration. Check which WDM platforms are IBM System z qualified and ensure that you are using the correct SFPs; otherwise, you can overdrive them in the WDM. You will need to configure the ISL ports as E-Ports (not auto-detect U_Ports) and use a fixed speed (not auto-negotiate). Also, the Long Distance Mode must to be configured for the correct distance. See 8.3.7, “Changing port type and speed” on page 179 and 8.3.8, “Changing buffer credits” on page 183 for configuration information. By default, the Director uses VC_Link_Init with ARB Fillword to initialize the ISL. However, some Dense WDM platforms do not support this mode. For those WDM platforms, you must use Idle Fillwords, or disable it and use ISL R_RDY Mode. Also, you must not use Quality of Service and Credit Recovery in this mode. They must be disabled before IDLE Fillwords are used, or ISL R_RDY Mode is enabled. ISL R_RDY does not support trunking. The following commands are needed: portcfgqos --disable slot/port to disable QoS. portcfgcreditrecovery --disable slot/port to disable credit recovery. portcfglongdistance slot/port LS 1 100 to enable Long Distance in LS Mode for 100 km with VC_Link_Init enabled for ARB Fillword (1). portcfglongdistance slot/port LS 0 100 to enable Long Distance in LS Mode for 100 km with VC_Link_Init enabled for IDLE Fillword (0). portcfgislmode slot/port, 1 to enable ISL R_RDY Mode (1). portcfgislmode slot/port, 0 to disable ISL R_RDY Mode (0). portcfgfillword slot/port 0 to set the Fillword to Idles (0). portcfgfillword slot/port 1 to set the Fillword to ARB(FF) (1). portcfgshow slot/port to verify the settings. Example 8-6 illustrates how to set up Long Distance Mode to 100 km (LS Mode = static) with ISL R_RDY Mode and Idles as Fillword, which is required for some WDM platforms.
Example 8-6 DWDM example for Port 1/15

portcfgqos --disable 1/15 portcfgcreditrecovery --disable 1/15 portcfglongdistance 1/15 LS 0 100 portcfgislmode 1/15, 1 portcfgfillword 1/15 0 portcfgshow 1/15

196

FICON Planning and Implementation Guide

8.6 FICON Directors in an intermixed environment
If you use FICON and Fibre channel in the same Director, we strongly recommend that you create two logical switches: one logical switch for FICON, and the other for Fibre Channel. For configuration details, refer to 8.3.3, “Setting up a logical switch (optional)” on page 170. This will allow you to configure a FICON logical switch with all required settings. and a Fibre Channel logical switch with its own settings.

8.7 Channel swap
A channel swap may be useful if you need to move a connected device to another physical port in the Director without changing the definition in the System z server or the zoning. The port will have the same address as it had before the swap. To perform a channel swap, right-click the director in the DCFM and select Element Manager → Ports. In the Port Administration window, click Show Advanced Mode in the upper right corner to see all possible functions. Disable both the port you want to swap and the port you want to swap to. Note: If you have intermittent errors to a System z server, set the attached CHPID offline before disabling the port on the Director side. Set the CHPID online after the swap is complete and the port is enabled again on the Director side. Select the port you want to swap to and click Port Swap. Enter the Slot Number and the Port Number of the “defective” port and click Swap. After the swap is complete, plug the cable to the new port in the Director. Enable the port after the cable is connected and check the Node Descriptor for a valid connection, as shown in 10.7, “Node descriptor” on page 226. Set the port online in z/OS.

8.8 Backing up Director configuration data
To save the Director configuration data, right-click any Director and select Configuration → Save. In the Save Switch Configurations dialog, select the Directors you want to back up and add them to the Selected Switches list by clicking the arrow pointing to the right. Click OK to save the Director configurations. You can also schedule a backup by right-clicking any Director and selecting Configuration → Schedule Backup. Make the desired settings in the Schedule Backup dialog and click OK. The backup will be performed automatically according to your settings. You can view the backups by right-clicking any Director and selecting Configuration → Configuration Repository. We recommend that you have a backup in case of a disaster (the configuration file can be restored in a replacement Director). You should also perform a backup before a firmware upgrade.

Chapter 8. Configuring FICON Directors

197

Click OK after the settings are done. In the Options window. Select the Restore tab and enter the location of the backup. The data can be restored via the DCFM Server console (located in the program directory). Therefore. Select the backup you want to restore and click OK. You will obtain a list of available backup files. right-click the Director you want to restore and select Configuration → Restore. Select SAN → Options from the DCFM menu bar. select Backup on the left and provide the necessary information. After you select the backup location. or browse the filesystem. Important: The restore of the Director’s configuration data is disruptive. click Restore. 198 FICON Planning and Implementation Guide .To restore. 8. you need to stop the DCFM services by clicking Stop. because the Director will perform a reboot to activate the new configuration. Note: We recommend that you save the backup on a network drive in order to be able to restore the data in case of a disaster or a hard disk failure.9 Backing up DCFM configuration data You should also backup the DCFM configuration.

All rights reserved. 2009. 2006. © Copyright IBM Corp.Part 4 Part 4 Managing the FICON environment This part discusses FICON operations and problem determination tools and techniques. Common issues are examined and descriptions are provided that explain how to resolve these issues using proven methods when working with FICON environments. 2005. 199 .

200 FICON Planning and Implementation Guide .

All rights reserved. 2006. 2005.9 Chapter 9. The following topics are discussed: System Activity Display (SAD) RMF Reporting DCFM performance monitor © Copyright IBM Corp. Monitoring the FICON environment This chapter discusses verification and monitoring tools and techniques for FICON environments. 201 . 2009.

1 System Activity Display The System Activity Display (SAD) shows individual system activity for System z servers (also known as Central Processor Complexes or CPCs). System activity includes the channel activity and physical processing activity that has been defined in the system activity profiles. Using the SAD. you can create a list of channels that you want to look at and monitor. leaving the channel waiting for a response but otherwise idle. RMF. if a performance problem is suspected to be caused by one or more FICON channels. activity will show a utilization that is significantly higher than that reported by RMF. Figure 9-1 SAD frame displaying high usage channel activity Note: The utilization reported by the Activity task for most channel types will correspond with the utilization reported by the Resource Measurement Facility (RMF).9. For Fibre Channels. or for a group of Central Processor Complexes. for performance troubleshooting. we recommend that you use the Resource Measurement Facility (RMF). which are stored in the selected CPCs. this task considers the channel to be busy any time an operation is pending. even if the channel is waiting for a device to respond. looks at the amount of work performed versus the amount of work that could be performed by the channel. This means that if you have devices that are relatively slow to respond. SC28-6830. 202 FICON Planning and Implementation Guide . in contrast. For more information about assigning and customizing activity profiles. however. However. Figure 9-1 shows a System Activity Display that is customized for high usage CHPIDs on processor SCZP201. refer to HMC Operations Guide.

The measurements we calculate to report I/O response time and that RMF reports on are: Queue time Pend time Connect time Disconnect time Figure 9-2 on page 204 indicates the I/O performance measurement points. and presents the data collected in the form of System Management Facility (SMF) records. For more information. An essential ingredient in performance monitoring is understanding the components of response time. it is imperative to establish your performance objectives.3 Introduction to performance monitoring Before you undertake any performance analysis.2 Resource Measurement Facility monitoring Resource Measurement Facility (RMF) is a performance management tool that measures selected areas of system activity.9.com/servers/eserver/zseries/zos/rmf/ 9. including different varieties of the FICON channel. It depicts a configuration with multiple servers exploiting a cascaded FICON Director environment.ibm. Chapter 9. This data is essential for any kind of FICON channel performance monitoring. Monitoring the FICON environment 203 . refer to the IBM RMF Web site at: http://www. without excessive tuning efforts. however. The goal of performance management is to make the best use of your current resources to meet your objectives. the measurement points are also valid for the other supported topologies (by excluding the PEND time for the second switch or both switches).

Connect time 6.1. PEND time ends at the FICON channel when the first command response (CMR) is received from the CU. Monitor II. RMF is comprised of several components: Monitor I. Disconnect time FICON DASD CU Figure 9-2 I/O performance indicators FICON exploits CCW chaining. PEND time (switch port activity) Switch 1 Switch 2 4. Also. PEND time (channel path activity) z10 EC Server z9 EC Server z10 BC Server 3. PEND time (CU activity) 5.4 Introduction to Resource Measurement Facility The Resource Measurement Facility (RMF) provides an interface to your System z environment that facilitates reporting and detailed measurements of your critical resources. so there will be fewer channel ends and device ends. Monitor III Postprocessor RMF Performance Monitoring Client/Server Reporter Sysplex Data Server Distributed Data Server LDAP back-end These components complement each other to provide the infrastructure for performance management: Gathering data Reporting data 204 FICON Planning and Implementation Guide . 9. Queue time (IOS queuing activity) z/OS 2 z/OS 1 z/OS 3 z/OS 4 z/OS 5 z/OS 6 2.

which can be invoked from either an ISPF dialog or directly with a TSO/e command (RMFMON) – A Monitor II background session. and overview reports In addition. the postprocessor can create overview records that are the optimal base for further spreadsheet processing on the workstation.4. The postprocessor is usually invoked via a batch job. or they can be generated at a later time by the postprocessor. which reflect a picture of performance for each interval for which the data has been gathered – Duration reports. which is a non-interactive session to create a report for printing RMF Monitor III runs as a TSO/e session under ISPF and provides sysplex or system performance reports by: – Displaying your current system status in real-time mode – Showing previously collected data that is still available in either storage buffers or preallocated VSAM data sets Monitor III offers a wide spectrum of reports for answering questions that arise during the various performance management tasks. RMF Monitor II is a snapshot reporting tool used to obtain very fast information about how specific address spaces or system resources (processor.Accessing data across the sysplex We introduce data gathering and RMF reporting in the next sections. which summarize data over longer periods of time with a maximum value of 100 hours – Summary. Monitoring the FICON environment 205 .2 RMF reporting RMF has three monitors and a postprocessor for reporting performance statistics: RMF Monitor I produces interval reports that are created at the end of a measurement interval. plot. All reporting is available within one TSO/e session. but it can also be run under a TSO/e session. 30 minutes. It offers the following types of reports: – Interval reports. but each monitor can be started sysplex-wide by one operator command. Chapter 9. for example. storage) are performing. DASD. You can obtain Monitor I session reports during or at the end of RMF processing. volumes.4. exception. 9. so there is no need to log on to different systems in the sysplex to get all performance data.1 Data gathering RMF gathers data using three monitors: Short-term data collection with Monitor III Snapshot™ monitoring with Monitor II Long-term data gathering with Monitor I and Monitor III The system operator starts all monitors as non-interactive (background) sessions with a variety of options that determine what type of data is collected and where it is stored. 9. The data gathering functions run independently on each system. Monitor II has two modes for reporting on system performance: – A Monitor II display session.

.0 0.000 .This is the average response time.000 .018 .002 .845 .0 0.000 . D320 DEVICE TYPE 33909 33909 33909 33909 33909 33909 33909 33909 33909 NUMBER OF CYL 10017 10017 10017 10017 10017 10017 10017 10017 600 IODF = 07 CR-DATE: 06/19/2009 DEVICE VOLUME PAV LCU ACTIVITY SERIAL RATE ACIV01 1.0H 0028 448.0 0. Example 9-1 DASD Activity Report 1 z/OS V1R10 TOTAL SAMPLES = STORAGE GROUP ACIRLS DEV NUM D308 D309 D30A D30D D30E D30F D310 D311 .0 0. and Connect times appear in a DASD I/O operation when a CACHE hit occurs.0 100. AVG DISC time .000 .05 % DEV UTIL 0.000 .0 0.000 .69 0.0 0.000 .0 1.000 .0H 0028 0.000 .000 #@$#X2 1.036 CYCLE 1.0 0.000 .0 100.00 0.000 .000 .0 100.000 IDGC00 1.000 .000 .000 SECONDS 43 0 10017 TBIG26 1.000 NWD30D 1.018 .0 0. AVG CONN time .000 .0 0.00.0 100.250 .0H 0028 0.128 .This delay is due to the device being busy because of I/O from another sharing z/OS system. Pend.000 .000 .000 .0H 0028 0.00 4.10 AVG AVG CMR DB DLY DLY .000 .00 0.0 0.0 100.000 .0 0.000 .0 0. – Channel path delay . It should be noted that with any RMF reports.000 .000 .000 .0 100.00 0.000 SAPTL3 1.1 DASD Activity Report The DASD Activity Report (see Example 9-1) is the first starting point for any I/O analysis and consequently one we should focus on in some detail..250 .5.000 .0 0.5 RMF example reports The following examples have been produced using the RMF postprocessor.0 100.000 .105 % DEV CONN 0.000 .0H 0028 0.This reflects the time when the device was in use but not transferring data.00.00 0.0 0.00 0.144 . in milliseconds.0 1.105 The fields of most interest are: AVG RESP TIME . control unit.0 D I R E C T A C C E S S D E V I C E DATE 06/23/2009 TIME 15.000 .000 .0 0.0 100.000 .000 .000 .000 .000 .000 ACT: ACTIVATE AVG AVG AVG PEND DISC CONN TIME TIME TIME .0H 0028 0. 206 FICON Planning and Implementation Guide .000 .0 0.0 0.0H 0028 0. If you are sharing I/O across multiple systems.000 .0 0.Any time that is not accounted for in the preceding two measurements is due to a delay at the SAP or channel.000 .000 .0 0.000 .000 .05 % DEV RESV 0. you will need to review reports from each of the sharing systems to see the complete I/O activity.0 0.000 .0H 0028 0.0 0.973 .000 .000 .000 10. the I/O information is reported from one z/OS system only.0H 0028 0.00 0.000 .63 0.000 .0 0.9. These delays could be caused by contention in the fabric and at the destination port.000 .000 COMST4 1.00 0.000 LCU 0028 448.00 0.144 .37 A C T I V I T Y PAGE SYSTEM ID SC33 RPT VERSION V1R10 RMF INTERVAL 10. for an I/O to this device or LCU.00 0.5 % % ANY MT ALLOC PEND 100.00 0.5 0.000 .002 .0 0. DASD) and transferring data.0 100.00 0.000 .000 .000 . Figure 9-3 on page 207 illustrates when and how IOSQ.39.000 .008 TBIG08 1.00 4. 9.00 0.This delay indicates the time between a successful I/O operation being initiated on the channel and the device acknowledging the connection.748 CR-TIME: AVG AVG RESP IOSQ TIME TIME .0 0.0 AVG NUMBER ALLOC 0.000 .00 0.000 FKDBU3 1. The measurement is composed of the sum of the following metrics: – AVG CMR (command response) Delay .000 .000 . or a CU busy.000 .0 0.00 0. – AVG DB (device busy) Delay .This is the time measured by the channel subsystem during which the device is actually connected to the CPU through the path (channel.780 . This can be considered as the overall measurement that reflects the health of the device.

Application I/O Request Start I/O (IOS-CSS) FICON DASD I/O Operation CU CACHE hit I/O Operation Ends IOSQ UCB Busy PAV/HiperPAV reduces it with extra UCBs Pend Channel Busy FICON Director Port Busy # of Buffer Credits CU/Device Busy Open Exchange limit Connect Working Transferring Data FC frame Multiplexing allows for better link utilization. more an awareness than a problem Connect Figure 9-3 FICON DASD I/O operation times with a CACHE hit Figure 9-4 illustrates when and how IOSQ. but may extend some connect times FICON connect time is not as predictable as ESCON.5. Pend. ESS Logical time. Application I/O Request Start I/O (IOS-CSS) FICON DASD I/O Operation CU CACHE miss / Extent conflict I/O Operation Ends IOSQ UCB Busy PAV/HiperPAV reduces it with extra UCBs Pend Channel Busy FICON Director Port Busy # of Buffer Credits CU/Device Busy Open Exchange limit Connect Cache miss ESS Logical Disconnect Connect Working Transferring Data Extent Confict No Reconnect Port Busy (buffer credits) Connect minus Logical Disconnect = Reported Connect Figure 9-4 FICON DASD I/O operation times with a CACHE miss or an Extent Conflict 9.2 I/O Queueing Report The I/O Queueing Report highlights percentages of activity rates and queue lengths for I/O components grouped by Logical Control Unit (LCU). Monitoring the FICON environment 207 . and Connect times appear in an DASD I/O operation when a CACHE miss or an Extent conflict has to be resolved by the control unit. A sample report is shown in Example 9-2 on page 208. Chapter 9.

000 0.00 0.00 0.00 0.105 11416.00 1451.00 0. and initially defined number of DCM channels for each LCU in this interval.245 0.00 0.00 1.42 1452.1 0.00 0.0 0.00 0.00 0.0 0.525 0.00 0.0 0.00 0.0 0.405 1538.00 0.0 0.060 0.0 0.0 0.127 0. – IOP UTILIZATION shows fields to measure the I/Os started and interrupted on each IOP and an overall percentage busy indicator.0 * 0.0 0.00 0.317 0.013 0.127 0.00 0.00 0.0 0.00 0.0 0.000 0.00 0.0 0.0 0.33 1451.0 0.00 0.490 2161. Reasons for each: • • • • CP .0 0.127 0.0 0. The rest of the report applies to I/O activity for the z/OS system being measured by RMF.0 0.00 2.0 0.00 0.133 0.00 0.580 0. maximum.881 0.Example 9-2 I/O Queuing Activity 1 I/O Q U E U I N G A C T I V I T Y SYSTEM ID SC30 DATE 06/23/2009 INTERVAL 10.00 0.00 0.0 0.0 0.00 0.81 8716.0 0.535 0.00 0.00 0.0 20A0 03 0.00 0.0 0.127 0.00 0.00 0.013 0.00 0.0 0.00 0.00 1453. – The Dynamic Channel Path Management (DCM) GROUP reports the minimum.00 0.00 0.00 0.37 CYCLE 1.Device busy z/OS System activity fields from the I/O Queuing Report – LCU and CONTROL UNITS show the logical assignment to the physical resources.00 0.267 0.127 0.10 ACT: ACTIVATE .399 0.127 0.00 0.00 z/OS V1R10 -TOTAL IOP 0 00 01 02 03 04 05 0 SYS LCU 0 0001 0 0002 0 0003 0 0004 0 0007 0 0008 0 000B 0 000C In a PR/SM™ environment (such as our configuration).0 0.00 0. Average queue length reflects the average number of entries on the initiative queue for each IOP.0 0.00 0.00 1.0 0.00 0.0 0.0 0.0 * 0.00 0.000 0.405 0.0 0.33 1453.535 1338.0 * 0.00.00 0.00 AVG AVG DELAY AVG AVG DATA CU DCM GROUP CHAN CHPID % DP % CU CUB CMR CONTENTION Q CSS HPAV OPEN XFER MIN MAX DEF PATHS TAKEN BUSY BUSY DLY DLY RATE LNGTH DLY WAIT MAX EXCH CONC 2080 02 0.127 0.IOP UTILIZATION -------.00 0.00 0.0 0.00 0.000 0.0 0.397 2342.00 0.0 4B00 4C 0.317 0. The fields we focus on are: PR/SM fields from the I/O Queuing Report – INITIATIVE QUEUE: This queue reflects the I/O initiation activity for each I/O processor (IOP).% I/O REQUESTS RETRIED --------.17 1452.00 0.00.0 * 0. – I/O REQUESTS RETRIED reflects the ratio of retried I/O requests.00 8716.000 0.0 0. Retries/SSCH (start subchannel) reflects the retried I/O requests initially started.00 0.00 0.133 0.0 0.00 0.00 1.Director port busy CU .00 1.00 0.0 0.00 0.00 0.105 0.0 0.0 0.127 0. The top section reports PR/SM system activity. the report is split into two sections.00 0.0 * 0.INITIATIVE QUEUE ------.00 0.00 1452.0 20E0 05 0.490 0.0 0.0 0.0 0.0 * 0.0 0.00 0.0 0.0 0.127 0.000 0.00 0.000 0.0 3080 1C 0.47 1451. 208 FICON Planning and Implementation Guide .0 0.00 0.0 * 0.0 30A0 1E 0.000 SECONDS SAMPLES = 600 IODF = 07 CR-DATE: 06/19/2009 CR-TIME: 10.00 0.037 RPT VERSION V1R10 RMF TIME 15.0 0.0 0.00 0.000 0.0 0.00 0.127 0.0 0.0 0.00 0.754 0.00 0.0 0.RETRIES / SSCH --------ACTIVITY AVG Q % IOP I/O START INTERRUPT CP DP CU DV CP DP CU DV RATE LNGTH BUSY RATE RATE ALL BUSY BUSY BUSY BUSY ALL BUSY BUSY BUSY BUSY 1454.Control unit busy DV . – CHPID taken shows how evenly I/O requests are spread across the available paths to the LCU.536 0.57 0.0 0.00 4B08 4E 0.00 1.525 2460.00 0.0 20C0 04 0.00 0.00 0.0 0. Note: IOP or I/O Processors refer to the System Assist Processors (SAPs) on your server.00 3.00 0.0 0.1 0.39.00 0.0 0.0 0.0 0.0 * 0. The activity rate reflects the assignment of I/O requests.14 1454.00 0.00 0.00 1452.0 0.754 1575.00 0.0 0.0 0.00 0.0 0.0 0.00 0.00 1451.Channel path busy DP .

UTILIZATION – PART reflects the utilization for this partition for this interval.35 0.5 1.30 0.10 ACT: ACTIVATE MODE: LPAR CPMF: EXTENDED MODE PATH UTILIZATION(%) READ(MB/SEC) WRITE(MB/SEC) FICON OPERATIONS ZHPF OPERATIONS G SHR PART TOTAL BUS PART TOTAL PART TOTAL RATE ACTIVE DEFER RATE ACTIVE DEFER OFFLINE 4 Y 9.01 0.00 0.77 2.037 CYCLE 1.0 86.0 4 Y 9.3 1. TYPE reflects that the type (FC_S) is FICON switched.96 3.– The percentage busy fields refer to Director Port (DP) and control unit (CU).0 0.66 1.0 0.01 0.22 1178.00 0.09 0.0 1. – CONTENTION RATE indicates the rate that the I/O processor places the request on the control unit header for this interval.0 0.16 0.3 Channel Path Activity Report Example 9-3 reflects a subset of a Channel Path Activity Report.15 0. – AVQ CSS DLY reflects a time (in milliseconds) that an I/O is delayed from the acceptance of a start on the subchannel to the point that the channel subsystem actually initiates the operation on this LCU.0 0.39 15.00 0.0 0.06 157.0 0.30 0.1 0.30 0.5 0.0 CSSID: 2 C H A N N E L SYSTEM ID SC30 RPT VERSION V1R10 RMF P A T H A C T I V I T Y PAGE DATE 06/23/2009 TIME 15.0 0.0 0.0 4 Y 0.67 1.14 0.51 0.5 0.000 SECONDS 2 The fields of interest are: CHANNEL PATH – – – – ID reflects the CHPID number.0 0.0 4 Y 2.0 4 Y 9.01 0.0 4 Y 1.20 0.75 1.96 3.58 3.1 0.0 4 Y 1.23 1178.8 1.58 0.0 0.0 OFFLINE OFFLINE 4 Y 9.23 0. – DELAY Q LENGTH reflects the average number of delayed requests on the control unit header for this interval.37 15.24 1176.01 0.5. G reflects the generation type (2 denotes a 2 GBps channel .7 1.96 3.35 0.1 1.0 4 Y 0.0 0.3 1. – AVG CMR DLY indicates the time (in milliseconds) between a successful I/O operation being initiated and the device acknowledging the connection. The value reflects the ratio of the number of times an I/O request was deferred due to either resource being busy. Example 9-3 Channel Path Activity Report 1 z/OS V1R10 IODF = 07 CHANNEL ID TYPE 51 FC_S 52 FC_S 53 FC_S 54 FC_S 55 FC_S 56 FC_S 57 FC_S 58 FC_S 59 FC_S 5A FC_S 5B FC_S 5C FC_S 5D FC_S 5E FC_S CR-DATE: 06/19/2009 CR-TIME: 10. 9.00.39.0 0.1 0.46 4.0 67.30 0.09 16.36 0. – TOTAL reflects the utilization from the entire processor for this interval.01 0.00 0.0 0.0 0.6 1.12 21.29 0.8 1.05 221. – AVG CUB DLY indicates the time (in milliseconds) that an I/O operation was delayed due to the control unit being busy.30 0.77 1.4 1.07 0.33 0. Monitoring the FICON environment 209 .36 0.0 1.1 1.40 0.37 INTERVAL 10.01 0.0 90.13 0.05 0.0 0.3 1.FICON).0 4 Y 0.0 0. – BUS reflects the percentage of cycles the bus was found to be busy for this channel in relation to the potential limit.71 0.41 15.23 1176.08 18.00.1 0.66 1. SHR indicates whether the channel is shared between one or more partitions.65 3.45 0.28 0.1 0.76 1.0 0.17 1.0 4 Y 2.5 0.00 0.15 0. Chapter 9.16 219.5 0.05 0.75 1.0 0. This occurs when all paths to the subchannel are busy and at least one path to the control unit is busy.13 1.4 1.0 0.14 153.01 0.97 3. The purpose of this report is to reflect channel path activity for the requested interval.01 0.0 0.0 0.0 0.0 0.0 0.39 15.4 1.63 1.0 0.0 0.1 0.0 65.25 0.20 0.18 0.

000 SECONDS CR-TIME: 10.01 0.01 0.11 0.5.14 0.WRITE -COUNT 1 56 72 388 796 683 71 94 973 747 72 397 106 455 68 0 957 0 56 279 965 329 459 107 40 284 397 449 934 76 686 171 0 111 0 0. 9. – ACTIVE is the average number of native FICON or zHPF operations that are concurrently active.00 0. and UTILIZATION displays for FICON channels.27 0.READ -.04 0.37 CYCLE 1. WRITE.” – DEFER is the number of deferred native FICON or zHPF operations per second.00 0.00 0.01 0.00 0.01 0.00 0.00 0. Often referred to as the number of “open exchanges.00 0. This is the number of operations that could not be initiated by the channel due to lack of available resources.4 FICON Director Report The measurements in this report reflect statistics regarding the ports that are performing I/O for all systems. but with only one LPAR defined.02 0.037 RPT VERSION V1R10 RMF TIME 15.00 0.17 0.00 0.00 0.05 0.00.10 ACT: ACTIVATE 62 TYPE: 006064 MODEL: 001 MAN: MCD PLANT: 01 SERIAL: 0000000119D2 AVG FRAME SIZE PORT BANDWIDTH (MB/SEC) ERROR READ WRITE -.-.00 0. – TOTAL shows the data transfer rate in MBps for this channel from the entire processor to the control unit for this interval.10 0. Example 9-4 FICON Director Report 1 z/OS V1R10 0 0 IODF = 07 CR-DATE: 06/19/2009 SWITCH DEVICE: 0062 SWITCH ID: PORT -CONNECTIONAVG FRAME ADDR UNIT ID PACING CU 8400 CU 8200 CU 8000 0D SWITCH ---0 0E CHP 86 0 0F CHP 54 0 10 CU ---0 11 CU ---0 12 CHP 87 0 13 CU ---0 14 CU ---0 15 CU ---0 16 CHP 8C 0 17 CHP 55 0 18 SWITCH ---0 19 CU ---0 1A CHP 8D 0 1B CHP 45 0 1C CU ---0 1D -----.00 0. Example 9-4 provides detailed configuration about each FICON Director and information pertaining to the connectivity and activity of the ports. WRITE(MB/SEC) – PART shows the data transfer rate in MBps for this channel from this partition to the control unit in this interval.07 0.READ(MB/SEC) – PART shows the data transfer rate in MBps for this channel from this partition to the control unit in this interval.00 0.---0 F I C O N D I R E C T O R A C T I V I T Y PAGE SYSTEM ID SC30 DATE 06/23/2009 INTERVAL 10.02 0. the PART columns of this report will reflect a zero (0) value for the READ.17 0. FICON and zHPF Operations (Physical Channel) – RATE is the number of native FICON or zHPF operations per second at the physical channel level.00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 210 FICON Planning and Implementation Guide .39. Note: On a machine running in LPAR mode.07 0. – TOTAL shows the data transfer rate in MBps for this channel from the entire processor to the control unit for this interval.00 0.00 0.00 0.00.00 0.

000 SECONDS 0SERIAL NUMBER 00000BALB1 TYPE-MODEL 002107-922 CDATE 05/13/2009 CTIME 21.8 ECKD WRITE 745.0 0. ERROR COUNT .1 0.1 00243 Undefined NO DATA TO REPORT OR ZERO 00300 FIBRE 2Gb ECKD READ 2.4M 618.1E 1F 20 21 CHP CHP CHP SWITCH 8E 95 42 ---- 0 0 0 0 72 0 0 56 169 0 0 58 0.The average frame size (in bytes) used to transmit and receive data during this interval.3 0.00 0.0 65.4K 1.00 0 0 0 0 Fields of significance are: AVG FRAME PACING .8K 5.36 CYCLE 1.237 RPT VERSION V1R10 RMF TIME 21.3 00232 FIBRE 2Gb NO DATA TO REPORT OR ZERO 00233 FIBRE 2Gb NO DATA TO REPORT OR ZERO 00240 FIBRE 2Gb ECKD READ 168.49. AVG FRAME SIZE READ/WRITE .1 ECKD WRITE 661.6 00241 FIBRE 2Gb ECKD READ 45. 9.6 -----72. Example 9-5 reflects the activity on a selected adapter.5 ESS Link Statistics Report This report should be used to analyze the external link usage and capacity planning of Peer-to-Peer-Remote-Copy (PPRC) links.00 0. PORT BANDWIDTH READ/WRITE .1K 12.3 0.3K 42.1 1. Example 9-5 ESS Link Statistics 1 z/OS V1R10 E S S L I N K S T A T I S T I C S PAGE SYSTEM ID SC31 DATE 05/13/2009 INTERVAL 10.00 0.0 0.0 0.8 00203 FIBRE 2Gb ECKD READ 899.4K 3.5 0.8K 10.9 00242 FIBRE 4Gb ECKD READ 5.The rate (in MBps) of data transmitted and received during the interval.2K 174.1 0.4 ECKD WRITE 9.7 -----7.0 ECKD WRITE 857.0 0. this report contains statistics by link type.4K 9.9 0.1 9.4K 58.7 00231 FIBRE 2Gb ECKD READ 2.5K 11.4 0. For each adapter (SAID) of the ESS.7K 2.1 0.9K 49.2K 20.8K 12.2K 72.0 3799.8K 78.0 63. Monitoring the FICON environment 211 .1 0.1 -----11.0 0.1 ECKD WRITE 142.1 2.7 -----72.6K 17.00 0.00 0.2 0.7 -----22.1 0.4K 5.00 0.0 -----0.2 8.1 ECKD WRITE 655.1 9.49.1M 553.1 7.6M 676.5K 2.1 0.2 6.3K 37.37 CINT 10.8 0.7M 5.2 3843.00 0.5 0.4K 315.8 ECKD WRITE 83.The number of errors that were encountered during the interval.5.2 0.7 ECKD WRITE 633.0 2 Chapter 9.This is the average time (in microseconds) that a frame had to wait before it could be transmitted due to no buffer credits being available.8 3808.9 00230 FIBRE 2Gb ECKD READ 2.6K 2.00 0------ADAPTER-------LINK TYPE-BYTES BYTES OPERATIONS RESP TIME I/O SAID TYPE /SEC /OPERATION /SEC /OPERATION INTENSITY 00201 FIBRE 2Gb NO DATA TO REPORT OR ZERO 00202 FIBRE 2Gb ECKD READ 1.9K 0.00.5 -----0.1 19.4 1.2 -----77.2K 1.0 67.0 3.

Decrease disconnect. if it is a suitable cache candidate. Use Dynamic Cache Management (DCME) to manage cache for you. A value of 1000 indicates that the link was busy for the entire time period. RESP TIME/OPERATION reflects the average response time (in milliseconds). connect. 9. BYTES/OPERATION reflects the average number of bytes transferred for each individual operation. Tune cache. if it is a suitable cache candidate. 212 FICON Planning and Implementation Guide . DASD Here we discuss basic practices for DASD. If necessary.6 General performance guidelines The following guidelines are basic practices that can be used to improve the performance of your FICON devices. LINK TYPE will be one of the following: – PPRC send and receive – SCSI read and write BYTES/SEC reflects the average number of bytes transferred for all operations on this link. if it is a poor cache candidate. Move or copy data sets to other volumes to reduce contention. Review blocksize.The fields are: ADAPTER – SAID reflects the System Adapter Identifier address. OPERATIONS/SEC reflects the average number of operations per second. Be aware that their applicability may vary based on vendor specifications and model type. – TYPE reflects the characteristics of the connection (2/4 Gb fiber. manually tune cache by turning off poor cache users. in our example). Un-cache the volume. Tune cache. High IOSQ time The following points reflect typical approaches to reduce IOSQ time: Implement Parallel Access Volumes. Add paths (if not already 8-path). I/O INTENSITY is measured in milliseconds/second. High disconnect time The following points reflect typical approaches to reduce disconnect time: Cache the volume. Decreasing the other response time components will also decrease IOSQ time proportionally. It reflects the duration of the interval for which the adapter was active. Use faster devices. A value greater than 1000 would indicate that concurrent operations had been active. Cache the volume. and PEND times. Increase where applicable. Reduce arm movement via data set placement (seek analysis required).5. Used Dynamic Cache Management (DCME) to dynamically manage cache for you.

In a cached environment. Allocate data sets to DASD. to as high as 20 (but no more than 20). with Automated Cartridge Loader). Reduce mount delay (for example. Exploit multiple allegiance if available on your DASD subsystem. increase where applicable. Chapter 9.5. Use VIO for temporary data sets. If you can. and thereby significantly reduce tape mount requirements. Check channel utilization. you can analyze the data flow from end-to-end or for selected ports.6 DCFM performance monitor With the DCFM performance monitor. High PEND time The following points reflect typical approaches to reduce PEND time: Change the mix of data on the volume to reduce contention. 9. Use PDSEs if you suspect directory search time is a problem. identify one data set that is contributing to most of the problem. Tape mount management (TMM) is a method that you can use to accomplish this. This may be eligible to be moved to a custom volume or moved into storage. Move or copy data sets to other volumes to reduce contention. Monitoring the FICON environment 213 . Use enhanced-capacity cartridges and tape drives. These are basic recommendations. Increase host buffers for priority jobs. Add channels (if not already one channel per control unit function).7 Tape The following list items reflect typical approaches to improving tape performance. The performance monitor must be configured and it consists of an End-To-End Monitor and a Real-Time Monitor. Increase buffers from the default of 5.Use I/O priority queuing (IOQ=PRTY in IEAIPSxx). and you should refer to the specific performance recommendations of your chosen solution for more guidance. Low channel utilization with high pend times means you might have to add more CU host adapters. to allow an application to use multiple drives). Implement Parallel Access Volumes. 9. especially if not a small number of channels per controller. if it is a suitable cache candidate. The significant developments in tape technology (such as the IBM Virtual Tape Server) will mitigate many of these points. Run requests sequentially to reduce contention. Review blocksize. Use indexed VTOCs. Check CMR time and estimate CU host adapter utilization to see if CU is being overloaded. High connect time The following points reflect typical approaches to reduce connect time: Cache the volume. use faster channels. Restructure for more parallelism (for example.

Select Monitor → Performance → End-to-End Monitors from the DCFM menu bar. Select the port on the left and click the arrow pointing to the right to add the port to the monitored list.6. 214 FICON Planning and Implementation Guide .1 End-to-End Monitor To set up an End-to-End Monitor. In the Set End-to-End Monitors dialog. Figure 9-5 End-to-End Monitor setup 3. click the arrow pointing to the right. as shown in Figure 9-6 on page 215. select on the left side of the window an initiator port (from the upper window) and a target port (from the lower window) as shown in Figure 9-5. To create the monitor pair. 6.6. 2. The End-to-End Monitor is now set up and can be used. 9. follow these steps: 1. follow these steps: 1. Click Apply to save the changes. 5. Now you can select the End-to-End monitor from the Monitored Pairs part of the window and select the Real-Time Graph or Historical Graph button. You can also select more measurements from the Additional Measures drop-down menu. 2. Select Monitor → Performance → Real-Time Graph from the DCFM menu bar. 4. You will get the realtime or historical data of the monitored device pair. Device Ports or ISL Ports).2 Real-Time Monitor To set up and use a Real-Time Monitor for some ports. Select the port type from the Show drop-down menu to view the ports displayed at the left part of the window (for example.9. 3.

The Real-Time Monitor is now set up and can be used.Figure 9-6 Real-Time Monitor setup 4. Click OK to save the changes. Chapter 9. You will get the real-time data of the monitored devices. You can also select more measurements from the Additional Measures drop-down menu. 5. Monitoring the FICON environment 215 .

216 FICON Planning and Implementation Guide .

and Printing program FICON Purge Path Extended Helpful z/OS commands Node descriptor DCFM logs DCFM data collection Common z/OS FICON error message . 2006. The following topics are discussed: Preparing for problem determination activities Environmental Record. 2005.interpretation © Copyright IBM Corp. 2009. All rights reserved. 217 .10 Chapter 10. Debugging FICON problems This chapter describes the problem determination (PD) tools and techniques that you can use in a FICON environment. Editing.

and control units.1 Using the D M . 10. In response to the D M=CONFIG(xx) command.1 Preparing for problem determination activities The problem determination (PD) information provided here assumes that the environment adheres to the architecture specification and implementation levels for the System z servers. 218 FICON Planning and Implementation Guide . the command D M=CONFIG(config-id) is used to show any deviation from the required operating configuration. more than one config-id may be needed to represent the I/O configuration at a particular time or interval. 10. remove. The D M=CONFIG(xx) command is used to call a member CONFIGxx in SYS1.1. the command can be added to member IEACMDxx in SYS1. they can simplify some problem determination tasks. and at the operator shift changeover to determine and be aware of any deviations from the initial configuration status. This command can be automatically performed during the IPL process. at a minimum). who can take the corrective actions.1.Display system Matrix command After your system has been IPLed. or recustomize a CU or device Dynamic I/O re-configuration activity for either software or hardware changes During scheduled I/O-related change activity. the system compares the contents of the CONFIGxx member and the current configuration. When a config-id member is created in the z/OS PARMLIB.PARMLIB to automate this process. The output of this command can be used later during problem determination to help you understand why the status of a hardware resource is in a certain condition.10. Example 10-1 on page 219 shows a CONFIGxx data set member. The D M=CONFIG(config-id) command is also useful after any of the following events: I/O-related repair activity I/O-related physical change activity – Changes to cabling – FICON Director settings – Add. customize. The CONFIGxx data set contains a predefined hardware configuration that should reflect all the resources critical to system operation. It then displays the differences (if they exist) to the operator.PARMLIB. the FICON Director. When the procedures discussed in this section are implemented in a z/OS z environment.2 Creating a CONFIG member A config-id member can be created by HCD from an existing IODF dataset. execute a D M command to create an entry in the z/OS system log that will function as a “footprint” to record the processor and I/O hardware configuration at IPL time. The D M=CONFIG(config-id) command can be invoked by Operations on a periodic basis (once every eight hours.

Problems since last Power-on Reset (POR). – This will cause the wrong IOCDS to be selected and used.ONLINE STOR(E=2). Example 10-2 D M=CONFIG(xx) output D M=CONFIG(S1) IEE097I 11.) – The wrong IODF device (volume) used may point to a different I/O configuration. Therefore. Examples of situations that have caused problems are listed here..Example 10-1 SYS1. or IPL Wrong HMC CPC reset profile used for the CPC for the Activate. I/O configuration definition changes performed that do not match the physical configuration.ONLINE .PARMLIB member. Wrong LOAD profile used. Wrong IOCDS associated to the Reset Profile used to activate.ONLINE DEV(0200-02FF).38.ONLINE STOR(E=0). If the current running configuration matches the definitions in the CONFIGxx SYS1. the output of the z/OS command would contain the message NO DEVIATION FROM REQUESTED CONFIGURATION.PARMLIB(CONFIGxx) CPU(0).ONLINE CHP(00-3F). – The wrong LOAD device used for the IPL may cause the wrong OSCONFIG to be used. Activate. Debugging FICON problems 219 . it is very important to be aware of all planned and unplanned changes. (The OSCONFIG member contains the z/OS software I/O configuration definition.13 DEVIATION STATUS 568 FROM CONFIGS1 DEVICE DESIRED ACTUAL D2A6 ONLINE OFFLINE D080 BASE HIPERPAV E200 OFFLINE ONLINE F890 PAV UNBOUND 10.1..ONLINE DEV(5100-6FFF). or wrong LOAD device or IODF device specified in the LOAD profile.ONLINE CPU(1). Example 10-2 shows the output of a D M=CONFIG(xx) command where deviations from a desired configuration (S1) exist.ONLINE STOR(E=1).ONLINE DEV(9A00-BF00). Chapter 10.3 Most common I/O-related problems Most common problems are due to changes made to the configuration.ONLINE CPU(2).ONLINE DEV(100O-2FFF).

2 Problem determination approach for FICON Before you initiate any problem determination tasks. “Using HMC and SE for problem determination information” on page 265. you need to have a clear understanding of the problem or condition. 10. FICON Director customization changes: Fabric.10. You also need current and accurate information and documentation to resolve problem situations. FICON Director port cabling changes. FICON Director zoning customization changes. or Port bindings. “Useful z/OS commands” on page 287 and in Appendix E. Determining the status of all components in the affected I/O path is also key. “Common z/OS FICON error message . Begin by taking note of IOS messages that are issued (see 10. FICON Director.interpretation” on page 230 for examples). Addition or removal of FICON Directors or control units. a significant amount of information is available in various areas. Both the z/OS commands and the HMC/SE problem determination facilities are discussed in Appendix F.Problems since the last physical configuration change Physical cabling changes at System z Server. The HMC and SE panels include specific Problem Determination functions which may also be required to use when investigating FICON channel problems. “Using HMC and SE for problem determination information” on page 265. The symptom or z/OS message associated with the problem that needs to be addressed The device number or device numbers of the problem device or devices The CHPID numbers of the problem channels or the channel paths to the device or devices The CPC and logical partition names that the z/OS is running on A configuration diagram of the complete I/O connectivity (all paths and FICON Director connections for a multi-path device) to the problem device The physical location of all involved hardware equipment and their respective consoles. DWDMs or Control Unit sides. IOS messages provide useful information and are an excellent starting point for problem determination. Problems since the last control unit customization changes Problems since the last FICON Director changes. FICON Director matrix changes (prohibit/allow). FICON Director port status changed: ports being blocked or taken offline. techniques to use for this task are shown in Appendix F. FICON Director. “Useful z/OS commands” on page 287 and Appendix E. when applicable For a FICON CTC-related problem: – Target CPC – Target Channel Subsystem – Target Image (LPAR) – Target matching device number (or the source/target CTC device pair) z/OS offers an extensive list of commands that will provide most of the required PD information listed. The following sections provide an overview of what can be used based on four 220 FICON Planning and Implementation Guide . patch panels. so try to have the information listed here available. For FICON channel problem determination.

levels of tools. FICON definition-related failures IOCP definition errors can cause initialization failures. Chapter 10. Fibers in general can cause initialization and IFCC failures. Debugging FICON problems 221 . Channel Detected Errors and Interface Control Checks (IFCC) can be caused by failures in FICON hardware or in the FICON definition setup. Control Unit definition errors can cause initialization and IFCC failures. The SE Problem Determination panels. shown under ‘“Displaying the Analyze Serial Link Status PD panel” on page 281. FICON hardware-related failures The FICON channel card can cause initialization IFCC and CCC failures. Fabric definitions can cause initialization and IFCC failures. initialization failures. z/OS commands for status z/OS system log and LOGREC EREP and Purge Path Extended HMC and SE provide: CHPID status CU and device status Node descriptor information SAD display z/OS level Problem Determination Approach System z Server Hardware level FICON Director FICON Director tools provide: Port status Node descriptor information logs FICON Control Unit Figure 10-1 Sources of information Storage tools provide: Port status Node descriptor information logs Most common FICON failures There are two common failures seen on FICON channels: Failure to complete initialization or establish a logical path Failure detected in the communications of command and data by either the channel or the control unit Initialization failures may or may not create a Hardware Message on the HMC/SE. The FICON I/O (CU) Host Bay Adapter (HBA) can cause initialization and IFCC failures. Figure 10-1 illustrates the different levels of software and hardware. will clearly show initialization failures on a per-FICON channel basis. as well as the tools provided from a verification perspective. Channel Control Check (CCC) errors. There is also a panel available under the Service logon user ID called “IFCC and Other Errors” that will display information related to initialization types of failures. The FICON Director can cause initialization and IFCC failures.

GA22-7589. Editing. all Fibre Channel (FC) channel paths can create extended subchannel logouts. or to learn about a particular error. selected software errors. under applicable error conditions. This information provides you with a history of all hardware failures.LOGREC. to determine the history of the system. The node descriptor information is gathered during the Fabric Login or Port Login when the logical path is established. as well as save it in the system log and in LOGREC. and generate I/O interruptions that indicate such a logout is pending retrieval. If a logout is pending for an I/O interruption. 10. which are referred to as Link Maintenance Information (LMI) records. 10. refer to z/OS MVS Diagnosis: Reference. IOS will retrieve the logout data and include the data in the subchannel logout record that it records in SYS1. For further information about this topic. the system records information about the error in the LOGREC data set or the LOGREC log stream. For additional information pertaining to LMI records. With FICON Purge Path Extended (PPE) enabled. z/OS can now register to receive FICON link incident records. Link Error Statistical Buffer (LESB) information is collected from each Fibre Channel port in the path from host to control unit.4 FICON link incident reporting To improve the ability to capture data for link error analysis. Use the Environmental Record.FICON zoning-related failures Zoning issues in fabrics can cause problems that are very difficult to diagnose. The channel subsystem 222 FICON Planning and Implementation Guide . and Printing program When an error occurs.5 FICON Purge Path Extended When the extended subchannel logout facility is installed and enabled. z/OS will display this information on consoles. refer to z/OS MVS Diagnosis: Tools and Service Aids. A link incident record consists of detailed node information that is gathered for particular link-degraded and link-failure incidents. If zoning has been implemented. GA22-7588. Note: A common zoning problem occurs when swapping or changing FICON channel cards and the channels are defined to a switch that has all the FCP. Check with your product vendor for details about this topic. 10. Editing. and selected system conditions. the zoning tables in the fabric must be updated whenever a new WWPN is installed and connected to a zoned fabric. FICON and Storage devices defined to their respective zones. Note: FICON Purge Path Extended must be supported by the control unit and FICON Director to receive error counters. and Printing program (EREP) to print reports about the system records. This information then results in the abort reason code being built.3 Environmental Record.

A subset of a sample Subchannel Logout Handler (SLH) record. SYS1. formatted by EREP.passes this information to z/OS as Extended Subchannel Logout Data (ESLD). The overall process is shown in Figure 10-2. and an SLH record is cut in the LOGREC medium. is included in Figure 10-3 on page 224.LOGREC SLH I/O interrupt IRB ESLD z/OS FICON .CU LESBs abort reason Figure 10-2 FICON Purge Path Extended error recording This function provides advanced problem determination of Fibre Channel bit errors.FC LESBs abort reason ESLD model-dependent Channel data Channel Abort reason code Channel port LESB Channel attached Port LESB FICON LESBs abort reason Director CUattached Port LESB data CU port LESB CU Abort error code CU model dependent data FICON . Debugging FICON problems 223 . Chapter 10.

SELECTIVE RESET SEQ CODE *** INVALID *** VALIDITY OF RECORDED DATA COUNT INVALID TERMINATION CODE VALID SEQUENCE CODE INVALID DEVICE STATUS INVALID CCW ADDRESS INVALID DEVICE NUMBER VALID SENSE DATA NOT STORED CHANNEL LOGOUT DATA 0000 00000000 00000000 00000000 0020 00000001 00000000 000023C7 0040 00000000 00613613 00006106 0060 01600CE6 50050764 00C1796A 0080 2200002A 00000000 02000000 00A0 00200100 30303231 30353830 00C0 00000000 00000000 00000000 00E0 00000000 00000000 00000000 CONTROL UNIT LOGOUT DATA 0000 00000000 00000000 0020 00000000 00000000 0040 00000000 00000000 0060 00000000 00000000 0080 00000000 00000000 00A0 00000000 00000000 --------------------------SCSW FLAGS--FLAG 0 FLAG 1 CCW FORMAT 1 RESERVED 0 PRE-FETCH CCW 0 SSCH FUNCTION 1 INIT STATUS 0 HSCH FUNCTION 0 ADDR LIMIT 0 CSCH FUNCTION 0 SUPP SUSPEND INT 0 RESUME PENDING 0 ZERO COND CODE 0 START PENDING 0 EXTENDED CONTROL 1 HALT PENDING 0 PATH NOT OPER 0 CLEAR PENDING 0 -- CODE 2 00000000 00000000 10000800 00190002 18100020 3049424D 00000000 00000000 00000000 0C000000 88A082F4 00000000 50050763 31333030 00000000 00000000 00000000 9000AB00 10000800 00000000 00CB945C 30303030 00000000 00000000 00000002 80000000 88A082F4 0800003E 50050763 30323232 00000000 00000000 00000000 80000000 50050764 00000000 00C0945C 31320024 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 Figure 10-3 Formatted Subchannel Logout Handler record Figure 10-4 on page 225 shows the CHANNEL and CONTROL UNIT LOGOUT DATA fields extracted from the EREP SLH record shown in Figure 10-3.DEVICE NUMBER: DEVICE TYPE: 0C07A 3390 REPORT: SCP: SLH EDIT VS 2 REL.TH TIME: 00 17 17. 3 DAY YEAR DATE: 275 08 HH MM SS. 224 FICON Planning and Implementation Guide .30 JOB CPU MODEL: 2084XA CHANNEL PATH ID: 57 LOGICAL CPU ID: 01796A PHYSICAL CHAN ID: 0131 PHYSICAL CPU ADDRESS: 31 CC CA FL CT FAILING CCW 00 00000000 00 0000 K FLAGS CA US SS CT 04 824017 00000008 00 02 0000 SUB-CHANNEL STATUS VOLUME SERIAL SUBCHANNEL ID NUMBER ERROR TYPE SCSW ---UNIT STATUS---- ATTENTION 0 PGM-CTLD IRPT 0 STATUS MODIFIER 0 INCORRECT LENGTH 0 CONTROL UNIT END 0 PROGRAM CHECK 0 BUSY 0 PROTECTION CHECK 0 CHANNEL END 0 CHAN DATA CHECK 0 DEVICE END 0 CHAN CTL CHECK 0 UNIT CHECK 0 I/F CTL CHECK 1 UNIT EXCEPTION 0 CHAINING CHECK 0 ----SOFTWARE RECOVERY STATUS----HARD FAIL 0 DEGRADE FAIL 0 SOFT FAIL 1 PASSED 0 CHANNEL ERROR ANALYSIS IRB STORED BY INTERRUPT TERMINATION BY -. Some fields are interpreted to give you an idea of how this information can be broken down.

(cc)) where dddd is the affected device address and cc is the failing CHPID – D M=SWITCH(xx) displays the ports configured and in use by the specified FICON Director number.1 where dddd is the affected device address DS P.t11. if issued from the affected logical partition (LPAR).dddd. refer to the ANSI standard architecture. can be very helpful for documenting the configuration and analyzing the errors in a FICON environment..6 Helpful z/OS commands The display results of the following commands. Also. before you begin to perform problem determination that involves a FICON Director. which can be found at: http://www. use these commands.CHANNEL LOGOUT DATA 0000 00000000 00000000 0020 00000001 00000000 0040 00000000 00613613 0060 01600CE6 50050764 0080 2200002A 00000000 00A0 00200100 30303231 00C0 00000000 00000000 00E0 00000000 00000000 CONTROL UNIT LOGOUT DATA 0000 00000000 00000000 0020 00000000 00000000 0040 00000000 00000000 0060 00000000 00000000 0080 00000000 00000000 00A0 00000000 00000000 Entry port on the LESB for the Director FICON Director entry port. obtain a copy of the z/OS syslog.org 10. Hex count of a specific error type 00000000 000023C7 00006106 00C1796A 02000000 30353830 00000000 00000000 00000000 00000000 10000800 00190002 18100020 3049424D 00000000 00000000 00000000 0C000000 88A082F4 00000000 50050763 31333030 00000000 00000000 00000000 9000AB00 10000800 00000000 00CB945C 30303030 00000000 00000000 00000002 80000000 88A082F4 0800003E 50050763 30323232 00000000 00000000 00000000 80000000 50050764 00000000 00C0945C 31320024 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 Channel F-Port LESB logout The Director entry port (Error counters are shown) B Error Code = 0C Indicates an ABORT was received from the Channel Channel N-Port LESB logout (clean in this example) Control Unit LESB counters (clean in this example) A D C FICON CHPID FICON Control Unit N-Port A x F-Port B D F-Port N-Port C Device C07A Figure 10-4 CHANNEL and CONTROL UNIT LOGOUT details For a detailed description regarding the debugging of these records. Debugging FICON problems 225 .. D M=CHP displays the status and type of CHPIDs D M=CHP(cc) where cc is the affected CHPID D U. Therefore. Chapter 10.dddd where dddd is the affected device address D M=DEV(dddd) where dddd is the affected device address D M=DEV(dddd.

. PORT STATUS 0 1 2 3 4 5 6 7 8 9 A B C D E F 0 u u u c u u u u u u u u u u u u 1 u u u u u u u u u u u c u u u u 2 u u u u u u u u u u u u u x u u 3 u u u u x u u u u u u u u u u u 4 . . . . . . serial number.7 Node descriptor The node descriptor will provide you with useful information such as the manufacturer. . . WWNN. . . . . . . Notice in the example that Switch 65. . . . . .15 DISPLAY M 215 SWITCH 0065. . . . . . . select Display and click the Set Up FICON Display button. . . . . . . . . along with their current state. . . 8 . . . . . . . these ports are used as ISL ports. . 7 . . node status. .7. 6 . . . .DCM NOT ALLOWED BY OPERATOR x NOT DCM ELIGIBLE p DCM NOT ALLOWED DUE TO PORT STATE c CHANNEL ATTACHED $ UNABLE TO DETERMINE CURRENT ATTACHMENT u NOT ATTACHED . . . .Examples of the z/OS commands can be found in Appendix F. . . . . . ***************** SYMBOL EXPLANATION ***************** + DCM ALLOWED . . Ports 2D and 34 display an x state. . . . . . . . . . . . . . . . . . . . . . F . 5 . . . . . DOES NOT EXIST The D M=SWITCH(xx) command provides important information about the configured and in-use ports at the specified switch. . Click OK to close the window. . . . “Useful z/OS commands” on page 287. . . Example 10-3 shows the output of a z/OS D M=SWITCH(xx) command. . . . . . . . .1 View node descriptors from DCFM If not already done. . . . . . B . . . . . C . refer to “Displaying the Analyze Channel Information panels” on page 272. . . . A . . . .45. . . . . . . . . . . . . . . 226 FICON Planning and Implementation Guide . you can change the view for FICON in the DCFM by clicking SAN → Options. . . . . . . . . . . Tag information. E . 9 . . . . . node type. ports 03 and 1B are in use and attached to a FICON Channel. . . Example 10-3 Output of the ‘D M=SWITCH(65) z/OS command D M=SWITCH(65) IEE174I 12. . . . For an explanation about how to obtain this information from a System z server. 10. . . 10. . D . . and WWPN. . . . . . . . . . . . . . . . . . In the Option window. . . . . . . . . . . . . . . . .

click Name Server at the left side of the window. Check mark the details you want to see. This will open the Name Server window displaying all connected devices. To change the port view.You will see the Node Descriptor details in the DCFM for each port at the left hand side of the window. You can also see the Node Descriptor for a Director by right-clicking and selecting Element Manager → Hardware. right-click a Director (on the left of the window) and select Port Display. as shown in Figure 10-5. as shown in Figure 10-6 on page 228. Figure 10-5 Node Descriptor overview Clicking the Accessible Devices button displays all devices that have access to the selected list entry (that is. Chapter 10. each device that is in the same Zone as the selected list entry). Selecting the list entry and clicking Detail View produces a detailed view for the selected device. Debugging FICON problems 227 . In the Directors Element Manager.

the Tag is a07e.1 and 3 0.Figure 10-6 Node descriptor detailed view The Tag field for the System z server is a four-digit number. by default. 228 FICON Planning and Implementation Guide . The Master Log is displayed in the lower part of the DCFM and it includes. as shown in Figure 10-7 on page 229. all events. You can also set a filter by clicking Filter and selecting the events you want show in the Master Log. which means that the CHPID is 7E and is defined in CSS 0 and 2 (see Table 10-1). select Monitor → Logs. If you want a specific log. Table 10-1 CSS mapping First digit 1 2 3 4 5 CSS 3 2 2 and 3 1 1 and 3 First digit 6 7 8 9 A CSS 1 and 2 1. and it consists of two parts. In our example.2 and 3 0 and 1 0.8 DCFM logs The DCFM has several logs that include information which will help you during problem determination.2 and 3 The Tag field details for a DS8000 are shown in “DS8000 port layout” on page 262. The last two digits are the CHPID and the first digit gives you information about the Channel Subsystem (CSS) in which the CHPID is defined.1.2 and 3 0 0 and 3 0 and 2 First digit B C D E F CSS 0.1 and 2 0. 10.

A new window will open. right-click the Director and select Telnet. To check the settings.9 DCFM data collection If you need to request service. One possibility is as follows: 1. Enter the command supportshowcfgenable ficon and press Enter. select Monitor → Technical Support → Collect Data in the DCFM menu bar. admin) and press Enter. Enter logout to close the window.) 2. setcontext 1 if you need to connect to a logical switch with FID 1. 6. (You can also use an ssh connection to the Director. you may be asked to collect data. Debugging FICON problems 229 . 10. Enter the user name (in our case. you must configure it for each Director. In the DCFM server or client. The Technical Support Data window appears. if needed. for example. To collect FICON-specific data. 5. To perform data collection.Figure 10-7 Selecting a Log You can display the log from a selected Director or logical Switch by right-clicking the Director and selecting Events. This is done via the command-line interface at the Director. 4. 3. Enter the password (the default is password) and press Enter. Chapter 10. Enter. as shown in Figure 10-8 on page 230. enter supportshowcfgshow.

“Most common I/O-related problems” on page 219.interpretation This section describes the most common z/OS FICON messages (IOS) and their respective meanings.10 Common z/OS FICON error message . Click the zip file and save it to your hard disk. and required actions when applicable. 230 FICON Planning and Implementation Guide . 10.1.3. explanations.1. the two major types of problems are discussed: Failure to complete initialization or establish a logical path Failure detected in communications of command and data by either the channel or the control unit This section explains the messages you may receive when encountering these problems.3\data\ftproot\technicalsupport\. If you are at the DCFM Server. to put it in the list of selected switches for data collection. Data collection is finished now and can be sent to the Support Center. Select the correct zip file by looking at the date and time stamp included in the file name (if the size is 0 KB. SA22-7637. For specific message details and actions. you can also go into the following directory to get the data: C:\Program Files\DCFM 10. After data collection is complete (which may take up to 30 minutes. In 10. depending on the size of your configuration). Read the message at the confirmation window and then click OK. Click OK to start the data collection. it means that data collection is not finished yet and you need to wait a few minutes longer). refer to z/OS MVS System Messages.Figure 10-8 Data collection window Select the Director or logical switch at the left side of the window and click the arrow pointing to the right. you can select the zip file by going to Monitor → Technical Support → View Repository.

For more information about Path Not Operational Mask.FORCE command. dddd represents the device in use. Refer to the explanations of messages IOS2001I or IOS2002I in this section for possible reasons why you might receive message IOS001E. Dynamic Pathing Switch Validation returns RC=12. refer to “Displaying the Analyze Subchannel Information PD panel” on page 275.CCC. or enter V dddd.FORCE to box the device. an altered FICON director configuration. generating a State-Change Ready interrupt. or a disabled channel path. IOS002A dddd.OFFLINE. APARs may also implement architectural changes regarding how messages are generated and displayed. and so on) was detected. The message text shows: dddd pp This was the device in use by an I/O operation when the failure occurred.CDC. On all paths that present Deferred Condition Code 3 (DCC=3). IOS001E dddd. POM was ON). depending on the z/OS release and version. which was possibly caused by power drop of the CU or FICON Director. Debugging FICON problems 231 . Tip: Check for a disabled Control Unit interface. Channel Data Check . The message text shows: dddd This was the device in use by an I/O operation when the failure occurred.. Channel Control Check . Deferred Condition Code 3).IFCC. This was the path in use by the device that become inoperative.> This message indicates that one or more paths went inoperative (indicated by the PNOM UCW field) which were previously operational (Path Operational Mask.OFFLINE. IOS050I CHANNEL DETECTED ERROR ON dddd.stat <PCHID=xxxx> This message indicates that a Channel Error (Interface Control Check . INOPERATIVE PATH pp<pp. The I/O requester will be kept waiting until: Path connection is physically restored (a V dddd. This section provides insight to the most common FICON problem reporting messages. Tip: Power up or Enable the Control Unit (or FICON Director) and then issue a VARY ONLINE command. Chapter 10.cmd. to indicate to the caller that the “no paths” bit is to be turned on in the device UCB and wait for the CU to become operational. We recommend that readers refer to the current z/OS Messages library for reviewing the output whenever a discrepancy is noted.ONLINE command may be required). PNOM and POM fields.pp.Note: The messages covered here may change. The device is BOXed by a V dddd. NO PATHS AVAILABLE This message indicates that all paths to the device were INOPERATIVE (DCC=3.

if known.stat <PCHID=xxxx> This message indicates that the Channel detected an error while the system was operating a device. This is the device and subchannel status. Tip: Run EREP. Various enhancements have been made via APARs OW47845 and OW57038 to provide more information about the possible error root cause. if known. Tip: The IOS will check the state of the timeout bit (the timeout indication is reported by the channel in the Extended Status Word Timeout (ESWT) before generating the IOS051I message for FICON channels.stat <PCHID=xxxx> This message indicates that the channel subsystem detected an interface timeout IFCC. Otherwise. if known. Otherwise. use the “IFCC and Other Errors” service task on the SE to further investigate possible reasons why this message was issued. Alternatively. then IOS050I will be issued instead. Both messages should be addressed at the hardware level. this field will be set to asterisks (*).chp. this field will be set to asterisks (*). This is the System z server Physical Channel IDentifier. This is the failing command being executed by the device. This is the channel path identifier (CHPID).op. The message text shows: dddd pp cmd stat PCHID: xxxx This was the device in use by an I/O operation when the failure occurred. IOS051I INTERFACE TIMEOUT DETECTED ON dddd.pp cmd stat PCHID: xxxx This is the channel path identifier (CHPID). this field will be set to asterisks (*). If timeout is not on. if known. Otherwise. Otherwise. Otherwise. This is the failing command being executed by the device. Otherwise. this field will be set to asterisks (*). this field will be set to asterisks (*). (Cross check is the most common Channel card hardware error). Note that IOS050I and IOS051I are the most common z/OS FICON error messages. or examine the SE CHPID work area for the IFCC Threshold Exceeded CHPID status. A Channel card hardware error has occurred. Otherwise. The message text shows: dddd pp This was the device in use by an I/O operation when the failure occurred. This is the channel path identifier (CHPID). if known.pp. This is the System z server Physical Channel IDentifier. IOS052I CHANNEL EVENT REPORTED FOR dddd. this field will be set to asterisks (*). 232 FICON Planning and Implementation Guide .cmd. this field will be set to asterisks (*). if known. if known. This is the device and subchannel status.

pp. this field will be set to asterisks (*). IOS071I dddd. MIH will find this condition and declare it as an IDLE WITH WORK QUEUED condition. START PENDING This message indicates that the Missing Interrupt Handler (MIH) detected one of these conditions: dddd.pp. if known. or in a thrashing state for paging I/Os).jobname. This occurs because the IMEX request cancels the pending sense operation so it can execute first. very busy with I/O interrupts being disabled. Refer to APAR OW50355. An IDLE WITH WORK QUEUED message may occur if an Immediate Execution (IMEX) Request is queued while a sense operation is pending for a device. Otherwise.jobname. Thus.jobname. CLEAR SUBCHANNEL INTERRUPT MISSING Tips: The jobname can be *MASTER*.pp. IOS075E dddd. Debugging FICON problems 233 . Channel Path pp may not be correct. To identify the correct path. the device gets MIH for IOS-issued I/O after 1 second (typically. Tip: This message indicates that the channel error has been recovered by the channel and it is logged only to SYSLOG (not to consoles). for further guidance. probably the system on the other end was not properly running (in a loop or wait condition. It is just the previous successfully selected path. if the default device class MIH is 00. For example. MISSING (CHANNEL AND) DEVICE END dddd. However. Look at the RMF report for the paging I/O rate. IDLE WITH WORK QUEUED dddd. if known. CTC devices). This is the System z server Physical Channel IDentifier. HALT SUBCHANNEL INTERRUPT MISSING dddd. and will be logged in SYS1. (No operator console message).00. this field will be set to asterisks (*).pp. Otherwise.jobname. In this case. refer to the EREP MIH record. IOS most likely is recovering I/O such as Unconditional Reserve (U/R) or Dynamic Path Selection (DPS) and the operation is generating the MIH condition. which corrects many MIH-related problems.pp. The message appears if the channel error has occurred with Extended Subchannel Logout Data pending and the Logout-only bit is ON.LOGREC.jobname. RECURRING MIH CONDITION FOR THIS DEVICE Chapter 10. IOS will not be able to start the IMEX request because the sense I/O is pending.cmd stat PCHID: xxxx This is the failing command being executed by the device. If Missing CE/DE for CTC devices. For Start-Pending. This is the device and subchannel status.

This message indicates that the Recursive MIH condition was not cleared by the previous MIH Clear Subchannel (CSCH) recovery sequence. As a result.jobname. or Clear Interrupt Missing). This means that the recursive MIH condition was detected before the original MIH recovery completed. CLEAR SUBCHANNEL INTERRUPT MISSING This message indicates that one of these conditions was detected: dddd.pp.ch.jobname.jobname.**. Or the TAPE MIH exit requested no CSCH to be issued when the current active I/O contains a long-running CCW. MISSING DEVICE END dddd. the original MIH message may not have been issued and this message was issued instead.pp. Tip: This message is issued when there is an active I/O operation that was not cleared by the CSCH recovery action sequence that was performed.jobname. Secondary Status Missing. change it to at least 2 seconds because the interval of the FICON Interface-Timeout IFCC (No Response from CU) is 2 seconds.pp.**. Halt Interrupt Missing. MISSING DEVICE END dddd.jobname.jobname. Tips: If you specify an MIH HALT interval that is too short (such as 1 second).pp. MISSING CHANNEL AND DEVICE END dddd. START PENDING 234 FICON Planning and Implementation Guide . such as Forward-Space File and so on.pp. IOS076E dddd. For a DASD device. Primary Status Missing. for example: The MIH was for CLEAR Subchannel interrupt missing. HALT SUBCHANNEL INTERRUPT MISSING The message indicates missing interrupt conditions where the MIH recovery did not reset the device (that is.**.jobname. there is no case where this occurs.jobname. CSCH may not complete quickly if the CSS-selected CHPID (based on the LPUM of the subchannel) is in a permanent busy condition. IOS077E dddd. CSCH was rescheduled due to the recursive MIH.jobname. Start Pending.ch. It is possible that the original MIH message may have been lost. IDLE WITH WORK QUEUED dddd. MISSING CHANNEL AND DEVICE END dddd. as indicated by the CONDITON flag being OFF in the MIH message module. The difference from message IOS077E is that this message is issued (instead of IOS077E) when the MIH message type is unknown (whether IOTIMING. CLEAR SUBCHANNEL INTERRUPT MISSING This message indicates that one of these conditions was detected: dddd. it did not issue the CSCH).jobname. START PENDING dddd.

IDLE WITH WORK QUEUED dddd. HALT SUBCHANNEL INTERRUPT MISSING Tip: This message means that another missing interrupt condition (MIH) was detected before the previous MIH recovery completed.dddd. LAST PATH pp LOST (text) This message indicates that one of these conditions was detected. where: dddd is the device and (text) is one of the following: RESERVE LOST CANNOT RE-RESERVE RESERVED ON FAILING PATH pp RESERVE MAY BE LOST MAY BE RESERVED ON PATH pp ASSIGN LOST CANNOT RE-ASSIGN WAS RESERVED WAS ASSIGNED STILL ASSIGNED STILL RESERVED ASSIGNED IN FAILING PATH pp Chapter 10. where (text) is one of the following: RESERVE LOST CANNOT RE-RESERVE STILL RESERVED RESERVE MAY BE LOST or MAY BE RESERVED ASSIGN LOST CANNOT RE-ASSIGN STILL ASSIGNED SELF-DESCRIPTION RECOVERY Tip: In the Channel Path recovery. Debugging FICON problems 235 . The device has been Boxed.jobname. This condition is generally associated to a hardware failure in the CU or device. the last path (CHPID) to the device has been varied offline. If other channel paths are available to the device. IOS100I DEVICE dddd BOXED.pp. vary them online.jobname. Any new I/O request to this device will result in Permanent I/O Errors.**. IOS101I DEVICE dddd FORCED OFFLINE {BOXED} (text) This message indicates that one of these conditions was detected.

RESERVE LOST C. This message itself may not show the reason of the device BOX. OPERATOR REQUEST (via the V dddd. IOS291I on all paths.WAS ASSIGNED STILL ASSIGNED .Reset Channel Path (System Reset). may be able to provide continuous availability to the device by swapping the faulty device (that is. Note that this condition could result from BOX_LP=(xxx) being specified for HOTIO or TERMINAL in the IECIOSxx PARMLIB member. or Reserve/Assign Lost in DPS Array.ALIAS UNBOUND DEFERRED BOX PROCESSING COMPLETE – The device had its BOX processing deferred and this message is issued to indicate that deferred BOX processing is now complete and the device is now in the boxed state. and so on) WAS RESERVED (BUT IT HAS BEEN RELEASED NOW) STILL RESERVED .4. It is usually accompanied by another IOS message (for example. the primary device) with a fully operational “backup” device (that is. IOS102I DEVICE dddd BOXED.I. BOX_LP=(ALL) is the default from z/OS 1. This message indicates that the system deferred the device dddd box processing to allow recovery processing to run. Tip: A permanent error on all Paths. REQUEST . where: dddd is the device and (text) can be one of the following.OFFLINE. (with ASSIGN or RESERVE LOST text). Any new I/O request to the device will result in a Permanent I/O error.MAY BE RESERVED . such as HyperSwap® operations. At the end of this process. See “IOS071I” on page 233IOS107I for more information. (text) This message indicates that one of these conditions was detected.FORCE command) PERMANENT ERROR (for example.U. IOS451I or IOS109E and so on) that may contain the reason for the device being boxed. Also.R. the secondary device) prior to permanently boxing the faulty device. an IOS102I message will indicate that the boxing process has completed. IOS107I DEVICE dddd BOX PROCESSING DEFFERRED Where: dddd is the device.NO PATHS .Tip: The Device Recovery routine invoked in the Channel Path Recovery detected one of these conditions as it tried to Re-Reserve/Assign the device after RCHP . be aware that if these messages are issued within the same time window (milliseconds). was detected during DPS Validation with message IOS451I. Tip: Certain recovery processing. This condition is generally associated to a hardware failure in the CU or device. 236 FICON Planning and Implementation Guide . then the SYSLOG message order may sometimes be out of sequence (reversed).

58 DISPLAY M 907 DEVICE D003 STATUS=ONLINE CHP 7E ENTRY LINK ADDRESS 6503 DEST LINK ADDRESS 6688 PATH ONLINE Y CHP PHYSICALLY ONLINE Y PATH OPERATIONAL Y MANAGED N CU NUMBER D000 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107.IBM. Tip: The message displays the STATUS FOR PATH(S) chp. Example 10-4 D M=DEV with information about inoperative paths D M=DEV(D003. IOS2001I dddd. Note that we entered the subject path in the DM command..0003 CHANNEL LINK LEVEL FACILITY DETECTED A LINK ERROR (90) LINK RCVY THRESHOLD EXCEEDED FOR ALL LOGICAL PATHS OF DEST LINK (60) HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW.Channel Path. IOS will retrieve the not-operational reason from the Channel Subsystem and issue message MSGIOS2001I (for MSGIOS001E or MSGIOS450E) or MSGIOS2002I (for MSGIOS002A).75. chp. When such a condition is present. the channel subsystem found an error before the system determined the device number.IBM. Temporary Error). While trying to service a request from a device.0000000BALB1. to SYSLOG only. contains the specific reason for the not-operative path (as shown in Example 10-4 on page 237)...0300 SCP TOKEN NED = 002107. UNSOLICITED MALFUNCTION INTERRUPT Where: pp is the path.(7E)) IEE174I 18. not-operational reason text. Check EREP. This message appears only in the SYSLOG. The second line of text.900. Debugging FICON problems 237 .0000000BALB1. Example 10-4 shows how the message text appears in the D M=DEV(dddd.0000000BALB1. the D M=DEV will report it via the text messages presented. INOPERATIVE PATHS Where: dddd is the device and pp is the path.(pp)) output.IOS162A CHPID pp ALERT.900.0000 SCP DEVICE NED = 002107. or MSGIOS450E is issued. NO ASSOCIATED SUBCHANNEL FOR DEVICE Where: pp is the path ID.900.75. MSGIOS002A.75.IBM. if it exists. IOS163A CHPID pp ALERT. with text describing the reason. (Channel Report Word received . ZHPF When message MSGIOS001E. Chapter 10.00.

The paths specified were found to be not operational for the specified device. Some FICON-relevant Status Code message texts are listed here: 50: CHANNEL SUBSYSTEM DETECTED A LINK FAILURE CONDITION) – 10: LOSS OF SIGNAL OR SYNCHRONIZATION CONDITION RECOGNIZED – 20: NOT OPERATIONAL SEQUENCE RECOGNIZED – 30: SEQUENCE TIMEOUT RECOGNIZED – 40: ILLEGAL SEQUENCE RECOGNIZED 60: CHANNEL LINK LEVEL FACILITY IN OFFLINE RECEPTION Note: This is the case when the Channel-to-CU logical path (H/W link-level) has been de-established (broken). this is the case where the CU logical path connection has been fenced because the up/down (flapping) error count has been exceeded. (Channel Path or Link level threshold exceeded). or Config CHPID OFF then ON (using a CF z/OS command or via the HMC). A0: LOGICAL PATH IS REMOVED OR NOT ESTABLISHED – 01: PACING PARAMETERS ERROR – 02: NO RESOURCES AVAILABLE – 04: DESIGNATED CONTROL UNIT IMAGE DOES NOT EXIST 238 FICON Planning and Implementation Guide . You must enter the V PATH (dddd. Note: For System z server FICON channels.pp). 70: PORT REJECT WAS ENCOUNTERED – 10: ADDRESS INVALID ERROR – 11: UNDEFINED DESTINATION ADDRESS ERROR – 12: DESTINATION PORT MALFUNCTION – 13: DYNAMIC SWITCH PORT INTERVENTION REQUIRED 80: LINK LEVEL REJECT WAS ENCOUNTERED – 01: TRANSMISSION ERROR – 05: DESTINATION ADDRESS INVALID ERROR – 07: RESERVED FIELD ERROR – 08: UNRECOGNIZED LINK CONTROL FUNCTION – 09: PROTOCOL ERROR – 0A: ACQUIRE LINK ADDRESS ERROR – 0B: UNRECOGNIZED DEVICE LEVEL 90: CHANNEL LINK LEVEL FACILITY DETECTED A LINK ERROR – 10: CONNECTION ERROR – 20: TRANSMISSION ERROR – 30: PROTOCOL ERROR – 40: DESTINATION ADDRESS INVALID ERROR – 50: DEVICE LEVEL ERROR (50) – 60: LINK RCVY THRESHOLD EXCEEDED FOR ALL LOGICAL PATHS OF DEST LINKS.ONLINE command to issue a Reset Link Recovery Threshold CHSC command.

This message appears only in the SYSLOG. Device Address 'dddd' is the trigger for Channel Path Recovery.– 05: LOGICAL PATH PRECLUDED BY CONFIGURATION AT CONTROL UNIT IMAGE – 06: LINK RECOVERY THRESHOLD EXCEEDED FOR LOGICAL PATH Note: For System z servers FICON channels. NO PATHS AVAILABLE Where: dddd is the device and pp is the path. IOS203I CHANNEL PATH pp SUCCESSFULY RECOVERED. Debugging FICON problems 239 . DEVICE IS: dddd / UNKNOWN Where: dddd is the device and pp is the path. B0: IN PROCESS OF INITIALIZING PATH • • • • • • 10: CONTROL UNIT DEVICE LEVEL INITIALIZATION IS NOT COMPLETE 20: LINK BUSY CONDITION LAST ENCOUNTERED 30: PORT BUSY CONDITION LAST ENCOUNTERED 30: PORT BUSY CONDITION LAST ENCOUNTERED 30: PORT BUSY CONDITION LAST ENCOUNTERED FF: NO FURTHER INFORMATION AVAILABLE OR UNKNOWN CONDITION C0: SUBCHANNEL PATH AVAILABLE. Path status code is created by Path State Info routine issuing Store Subchannel Path Information (SSPI). You must enter the V PATH (dddd. Channel Path Recovery was successfully completed. or Config CHPID OFF then ON (using a CF z/OS command or via the HMC). Note the following examples: FICON cables were disconnected at the CU side through FICON SW: STATUS FOR PATH(S) pp IN PROCESS OF INITIALIZING PATH (B0) NO FURTHER INFORMATION AVAILABLE OR UNKNOWN CONDITION (FF) FICON Director Power-Down (by power button): CHANNEL LINK LEVEL FACILITY IN OFFLINE RECEPTION STATE (60) IOS202E CHANNEL PATH pp FORCED OFFLINE. this is the case where CU logical path connection has been fenced due to up/down (flapping) error count exceeded. The CHPID has been configured OFFLINE from the Hardware Management Console (HMC). DEVICE IS: dddd / UNKNOWN Where: dddd is the device and pp is the path. or it was forced offline when Channel Path Recovery (RCHP) failed to clear the path in an error condition.ONLINE command to issue a Reset Link Recovery Threshold CHSC command. BUT DEVICE NOT OPERATIONAL IOS2002I dddd. Chapter 10. Tips: All paths to the specified device were found not operational.pp). (Channel Path or Link level threshold exceeded).

uses a Support Element (SE) component to request a config off or on of all the CSS.This is caused by the following conditions: CRW w/ Channel Path. called System Initiated CHPID Reconfiguration.CHPIDs associated with a particular channel ID. This facility.CHPIDs.. in a repair action. Terminal condition (I/F Hung) Reset-Event-Notification (Path Verification) Hot I/O & Action specified for its message is CHPK IOS288A SYSTEM-INITIATED OFFLINE | ONLINE RECONFIGURATION IS IN PROGRESS FOR THE FOLLWING CHPIDS: CC.cc-cc. With this facility. CC-CC. This message reports the results of reconfiguration actions requested and reported by IOS288A. IOS289I SYSTEM-INITIATED OFFLINE | ONLINE RECONFIGURATION HAS COMPLETED. The results can be: RECONFIGURATION WAS SUCCESSFUL FOR THE FOLLOWING CHPIDS: cc... Normally. 240 FICON Planning and Implementation Guide . the operator has to take the channels offline using the z/OS commands to configure off or on all of the associated CHPIDs in all partition.cc-cc RECONFIGURATION FAILED FOR THE FOLLOWING CHPIDS: cc. System z servers allow the use of the Hardware Management Consoles (HMCs) to “Configure CHPIDs On and Off” without “stealing” the channel from the configuration. with the exception of the last path to a device. Where: cc is the channel path. z/OS handles the SE requests configuring all the CSS.

Part 5 Part 5 Appendixes © Copyright IBM Corp. 241 . 2005. 2006. 2009. All rights reserved.

242 FICON Planning and Implementation Guide .

2009. Example: planning workflow This appendix shows an example of a workflow using the steps described in Chapter 4. “Planning the FICON environment” on page 59. All rights reserved. © Copyright IBM Corp. 2006. 2005.A Appendix A. 243 .

“Configuring a cascaded FICON Director topology” on page 133. That information can be used as a starting point for your documentation. Based on our physical and logical inventory. High availability is not a key requirement. these were the components involved in the solution: An IBM System z10 Enterprise Class in the main location – Two z/OS V1R10 LPARs – Two FICON Express8 channels An IBM System Storage DS8000 in the remote location – Two storage control units and two devices – Two 4 Gbps ports that will be upgraded to 8 Gbps in the near future Two IBM System Storage SAN b-type family (FICON Directors) – SAN768B with four 8 Gbps ports (in the main location) – SAN384B with four 8 Gbps ports (in the remote location) Figure A-1 displays the components used to create our new FICON environment.Documentation Creating and maintaining documentation throughout all the planning. Step 2 . and Chapter 8. design. Configuration worksheets for your FICON Director and FICON CTC environments can be found in Appendix B. “Configuration worksheets” on page 251.Requirements We based our requirements on the need to access non-business-critical data in a remote location from two z/OS LPARs in the main location. z10 EC SCZP201 LPAR A23 z/OS V1R10 FICON Express8 PCHID 5A2 System Storage SAN384B Port 88 Port 8D Port C4 Port CF DS8000 Port 0003 LPAR A24 z/OS V1R10 FICON Express8 PCHID 5D2 Port 03 Port 2D Port 34 Port 0242 Port 1B SAN768B Figure A-1 Results of the inventory effort 244 FICON Planning and Implementation Guide . “Configuring FICON Directors” on page 151. Isolating the new configuration from the existing one is essential to the solution. For documentation of our implementation steps for a cascaded FICON Director environment see Chapter 7. Throughout this book we provide information regarding the various features and functions offered by System z servers and the FICON Directors (and under which conditions they should be applied).Step 1 . and implementation phases is very important.

5. In Figure A-2.Context Taking into consideration the existing and planned components (System z. best fit our requirements. FICON Director. All ports will be running at date link rate of 8 Gbps. The transceiver type required for the ISLs in a cascaded FICON Director topology is long wavelength (LX). LX transceivers are generally recommended for all connections.4. Example: planning workflow 245 . Ports 0003 and 0242 in the DS8000 will be upgraded at a later point. which we used for all our ports. “Moving to a high bandwidth environment (FICON Express8)” on page 65 and 4. z10 EC DS8000 FICON Express8 LX LX LX SAN384B FICON Express8 LX LX LX LX LX LX LX LX SAN768B LX Figure A-3 Our cascaded FICON Director topology Appendix A. For more information about supported distances. The two locations were less than the maximum distance of 10 km apart. see Figure A-3. “Migrating from a single site to a multi-site environment” on page 66.2.4. ports 2D and 34 in SAN768B and ports 8D and C4 in SAN384B will be used for ISLs. refer to Table 4-1 on page 67.Step 3 . z10 EC SCZP201 LPAR A23 z/OS V1R10 FICON Express8 PCHID 5A2 DS8000 Port 03 Port 2D Port 1B Port 34 SAN384B Port 8D Port C4 Port 88 Port CF Port 0003 LPAR A24 z/OS V1R10 FICON Express8 PCHID 5D2 SAN768B Port 0242 Figure A-2 Our FICON components in the two locations Step 4 . and storage control device).Topologies and supported distances We implemented a cascaded FICON Director topology (see 4. “Cascaded FICON Directors” on page 68 for more information). therefore a WDM environment was not required. we determined that the scenarios 4.3. they will run at a date link rate of 4 Gbps. except for the ports in the DS8000.3.

1 Firewall Corporate Network System Storage SAN384B Figure A-5 Our FICON Director management environment 246 FICON Planning and Implementation Guide .1. we secured the fabric using zoning and the fabric configuration server with strict SCC policies. other relevant factors for the FICON fabric such as power consumption.1.21 10. the FICON Directors do offer this possibility.1. Our FICON environment was connected to an isolated network within a secured area. as illustrated in Figure A-4.1. Figure A-4 FICON management via DCFM Our FICON Director management environment is shown in Figure A-5. Data Center Fabric Manager Server/Client Data Center Fabric Manager Client System Storage SAN768B 10.1.1.1. To isolate the new configuration from the existing configuration. Because we were using existing environments.1. The figure also shows the IP addresses that we used for our FICON Directors. and firewall/router. DCFM server.Management For management purposes. However.Step 5 .1. Step 6 . we highly recommend that you change all the default passwords for all default accounts.77.77.77.1.61 10. if needed.1.1.1.1. therefore we only used username and password authentication and authorization.10 CP0 CP1 SW 66 10. cooling.1.62 10.77.20 SW 65 CP0 CP1 10.22 10.32 10.Convergence In our FICON environment we did not set up an intermix environment. we used the Data Center Fabric Manager (DCFM) for the entire FICON environment including setup and monitoring.30 10. and space inside the data center were not a concern. However.31 10.1.

We also used the CUP and SA for I/O to manage and control System z FICON connectivity.

Step 7 - Virtualization
In System z, we had two z/OS LPARs. Although the two z/OS LPARs use the same CSS, each FICON channels will be defined to an additional CSS (spanned) for future use. The channels were on two different FICON features. We defined redundant paths from both z/OS LPARs to the storage control unit.

z10 EC
SCZP201
LPAR SC30 A23 z/OS V1R10
FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E

System Storage SAN384B
Switch # 66 Switch @ 66
Port 8D Port C4 Port 88 Port CF

DS8000
CU# D000
LX 0003

D0xx

LPAR SC31 A24 z/OS V1R10

FICON Express8 LX CSS2 PCHID 5D2 CHPID 82

Port 03

Port 2D

Port 34

Switch # 65 Switch @ 65

LX 0242

D1xx

CU# D100
Port 1B

SAN768B
Figure A-6 Virtualization

For the DS8000, we set up two ports for redundant access to all devices. We had two logical control units inside the DS8000, and we defined the DS8000 channels to different FICON features. Our FICON Directors were, by default, set up with virtualization support and we used this technology to define port addresses beyond the FICON port addressing range. We created a new logical switch to be able address the 48-port line cards. We also assigned our ports to two different cards for redundancy. The two virtual FICON Directors were interconnected with ISLs. The following two figures show the worksheets that were used to configure the FICON connectivity for the z10 EC, the two FICON Directors, and the DS8000. Figure A-7 on page 248 displays the configuration worksheet for FICON Director 1.

Appendix A. Example: planning workflow

247

FICON Director Configuration Worksheet
FICON Director Manufacturer: IBM Type: 2499 Model: 384 S/N: 02-8D014

HCD Defined Switch ID x65 FICON Director Domain ID x65
FICON Director Ports
Slot Number Port Number Port Address Laser Type: LX / SX

(Switch ID) (Switch @)

Cascaded Directors No Yes x Corresponding Cascaded Director Domain ID x66 Fabric Name Cascaded Fabric 1

Attached N_Ports / E_Ports (CU, CPC, or ISL)
Port Name Node Type
CU/CHNL/FD

Machine Type

Model

Serial Number

ISL CU I/F CPC CHPID

1 2 3 4

3 11 13 4

6503 651B 652D 6534

LX LX LX LX

Z10 EC PCHID 5A2 Z10 EC PCHID 5D2 ISL1 to SAN384B ISL2 to SAN384B

CHNL CHNL FD FD

2097 2097 2499 2499

E26 E26 192 192

02-1DE50 02-1DE50 02-2C00Z 02-2C00Z

CPC CHPID CPC CHPID ISL ISL

Figure A-7 Configuration worksheet for FICON Director 1

Figure A-8 displays the configuration worksheet for FICON Director 2.

FICON Director Configuration Worksheet
FICON Director Manufacturer: IBM Type: 2499 Model: 192 S/N: 02-2C00Z

HCD Defined Switch ID x66 FICON Director Domain ID x66
FICON Director Ports
Slot Number Port Number Port Address Laser Type: LX / SX

(Switch ID) (Switch @)

Cascaded Directors No Yes x Corresponding Cascaded Director Domain ID x65 Fabric Name Cascaded Fabric 1

Attached N_Ports / E_Ports (CU, CPC, or ISL)
Port Name Node Type
CU/CHNL/FD

Machine Type

Model

Serial Number

ISL CU I/F CPC CHPID

7 8 7 8

8 15 13 4

6688 66CF 668D 66C4

LX LX LX LX

DS8K 10003 DS8K 10242 ISL1 to SAN768B ISL2 to SAN768B

CU CU FD FD

2107 2107 2499 2499

E26 E26 384 384

75-BALB1 75-BALB1 02-8D014 02-8D014

CU I/F CU I/F ISL ISL

Figure A-8 Configuration worksheet for the FICON Director 2

For high integrity, we set up Insistent Domain IDs and strict SCC policy. We used one port-based zone for all our FICON channels.

Step 8 - Performance
The short distance between the two locations allowed us to keep the default buffer credit assignment in the FICON Directors. For high performance, we used the zHPF feature in both the System z server and the DS8000. For better performance of the fabric links, we used lossless DLS with frame-based trunking. We also used port fencing to ensure that no loss of performance occurs due to excessive errors on the interfaces, and to make certain that the fabric remains stable.

248

FICON Planning and Implementation Guide

To monitor and evaluate the utilization of the paths and the traffic patterns, we used RMF on our System z server.

Step 9 - Prerequisites and interoperability
We ensured that all System z, DS8000, and FICON Director prerequisites and interoperability requirements are fulfilled, for example: System z10 – Feature code for FICON Express8 LX FC3325 – z/OS 1.7 with IBM Lifecycle Extension for z/OS V1.7 (5673-A01) DS8000 – Feature code for zHPF FC0709 and FC7092 SAN768B and SAN384B – Feature code for the FICON CUP management FC7886 See references in 4.10, “Prerequisites and interoperability” on page 92.

Step 10 - Physical connectivity
Our setup was based on LX FICON transceivers with standard LC duplex connectors, therefore we used 9 µm single-mode fiber optic cables (OS2 fiber type); see Figure A-9.

z10 EC
FICON Express8 LX PCHID 5A2 CHPID 7E

System Storage SAN384B
LX 88 LX 8D LX C4 LX CF

DS8000
SM

9µ m

9µ m

SM
LX 03

LX 0003
m 9µ SM

m 9µ SM
LX 2D LX 34

FICON Express8 LX PCHID 5D2 CHPID 82

m

SM
LX 1B

LX 0242

SAN768B
Figure A-9 Physical connectivity

All cable connectors are LC Duplex type

For the cabling infrastructure we used IBM Fiber transport service products with single mode fiber.

Conclusion
After all these steps were completed, our final FICON design appeared as shown in Figure A-10 on page 250.

Appendix A. Example: planning workflow

249

z10 EC
SCZP201
LPAR SC30 A23 z/OS V1R10
FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E

SAN384B
Switch # 66 Switch @ 66
Port 8D Port C4 Port 88 Port CF

DS8000
9µ m SM

CU# D000
LX 0003

9µ m

SM

D0xx

m 9µ SM
Port 2D

L IS

L IS

9µ m SM

LPAR SC31 A24 z/OS V1R10

FICON Express8 LX CSS2 PCHID 5D2 CHPID 82

Port 03

Port 34

9µ m

Switch # 65

LX 0242

D1xx

SM

Switch @ 65

CU# D100
Port 1B * All cable connectors are LC Duplex type

SAN768B
Figure A-10 Our final FICON design

250

FICON Planning and Implementation Guide

B

Appendix B.

Configuration worksheets
This appendix contains the following worksheets: FICON Director Configuration Worksheet Use this worksheet to document the layout of your FICON Director. It can be applied as a tool to help you understand how the ports are allocated for configuration and problem determination purposes. FICON CTC Image-ID Worksheet Use this worksheet as an aid in planning and documenting the association between LPARs, MIF IDs, Logical Channel subsystems, and CTC Image-IDs.

© Copyright IBM Corp. 2005, 2006, 2009, 2006. All rights reserved.

251

252

FICON Director Configuration Worksheet
Model: ____ S/N: _____________

FICON Director Manufacturer: ______________ Type: ______
Cascaded Directors No ___ Yes ___ Corresponding Cascaded Director Domain ID _______ Fabric Name ________________________________

HCD Defined Switch ID _____ (Switch ID) FICON Director Domain ID _____ (Switch @)

FICON Director Ports
Port Address Port Name Model Node Type CU / CHNL Machine Type Laser Type: LX / SX

Attached N_Ports / E_Ports (CU, CPC, or ISL)
Serial Number ISL CU I/F CPC CHPID

FICON Planning and Implementation Guide

Slot Number

Port Number

Server: Type: LPAR Name CSS MIF ID CTC-Image ID Server: Type: LPAR Name CSS MIF ID CTC-Image ID Appendix B. Configuration worksheets 253 .

254 FICON Planning and Implementation Guide .

2006. 2009. CMT. These statements and keywords are described in “IOCDS statements and keywords used for FICON” on page 257. and the devices have to be specified and saved in an I/O Configuration Data Set (IOCDS).C Appendix C. Also included are descriptions of statements and keywords that are used to define a FICON switched Point-to-Point topology. HCD. IOCP is a stand-alone program running in a partition on a System z server. HCM and CMT are PC-based programs. the control units. SG24-7571. with System z all paths to an I/O device must be defined (there is no auto-detection or auto-configuration). The IOCDS file is stored on the Support Element. Refer to IBM System z10 Enterprise Class Configuration Setup. The tools are: Hardware Configuration Definition (HCD) Hardware Configuration Manager (HCM) CHPID Mapping Tool (CMT) I/O Configuration Program (IOCP) HCD is a program that runs in a z/OS image. 255 . and device. HCM. control unit. The IOCDS contains statements and keywords which specify the characteristics of a channel. The characteristics of the channel paths. which is used by a System z server to load the configuration into its storage. Unlike other platforms. 2005. IOCP can create an IOCDS only for the physical system it is running on. and IOCP are briefly described in this appendix. © Copyright IBM Corp. for additional information about how to set up the System z server. All rights reserved. Configuration and definition tools This appendix provides an overview of tools that you can use to define the path between a System z server and a DASD control unit and a device. Although HCD and HCM can create an IOCDS for each physical CPC present in a sysplex environment.

for detailed information about HCM. most importantly. SB10-7037. I/O Configuration Program The I/O Configuration Program (IOCP) is a stand-alone program running in partition on a System z server.Hardware Configuration Definition The Hardware Configuration Definition (HCD) supplies an interactive dialog to generate your I/O definition file (IODF) and subsequently your Input/Output Configuration Data Set (IOCDS). you can also manage the physical aspects of a configuration. Additional enhancements have been built into the CMT to help with the requirements of z9 and z10 servers. Refer to HCD User’s Guide. and the best availability recommendations for installed features and defined configurations. Refer to HCM User’s Guide. SC33-7989.com/resourcelink The CMT program and the CMT users guide can be downloaded to a workstation from IBM Resource Link. The stand-alone IOCP can be used to manage and modify the input/output (I/O) configuration source files and data sets for the selected central processor complex (CPC) only. for detailed information about IOCP. to avoid connecting critical paths to single points of failure. Hardware Configuration Manager The z/OS Hardware Configuration Manager (HCM) is a PC-based client/server interface to HCD that combines the logical and physical aspects of hardware configuration management. SC33-7988. More details and informations about the CHPID Mapping Tool are available on IBM Resource Link™ at the following URL: https://www. as opposed to writing your own IOCP statements. All updates to your configuration are done via the HCM intuitive graphical user interface and. It is strongly recommended that you use HCD to generate your IODF. all changes of the logical I/O configuration are written into the IODF and fully validated and checked for accuracy and completeness by HCD. you can effectively manage the flexibility offered by the FICON infrastructure. 256 FICON Planning and Implementation Guide . It resides on the support element of a System z server and is loaded into an active partition when the Input/output (I/O) Configuration icon is clicked in the support element workplace. In addition to the logical connections. For example.ibm. Refer to IOCP User’s Guide. The validation checking that HCD performs as you enter data helps to eliminate errors before you implement your I/O configuration. CHPID Mapping Tool The CHPID Mapping Tool (CMT) provides a mechanism for customizing the CHPID assignments for System z servers. for detailed information about HCD. thereby providing a mechanism to map CHPIDS to PCHIDS.

The Switch keyword is required. It is recommended that you have the same value for the switch number and the switch address.com/resourcelink ID statement The ID statement is used to specify identification information of the server and IODF used to build the IOCDS. Optionally. only keywords used to define a control unit connected to a FICON channel (FC) are listed. In addition to other keywords allowed in the ID statement. only keywords used to define a FICON channel (FC) are listed. The partition names and their MIF IDs and associated CSSs are defined here. Configuration and definition tools 257 . The latest version of this publication is available on IBM Resource Link: https://www. Note that the switch number defined in the IOCDS is not the same as the switch address that is set up in the FICON Director. This indicates that the FICON channel will operate in FICON native (FC) mode.IOCDS statements and keywords used for FICON This section describes the required keyword definitions and rules for a FICON configuration. and it specifies an arbitrary number for the FICON Director to which the channel path is assigned. The LINK keyword is required to define the destination port address where the control unit is connected to the FICON Director. PATH The PATH parameter allows you to specify the CSSs and the CHPID number that will be assigned to a channel path. RESOURCE statement The RESOURCE statement provides information about logical partitions. CHPID statement The CHIPD statement contain keywords that define the characteristics of a channel. The PCHID keyword will be used to identify the location of the physical channel in the processor. only keywords used later in this chapter are described here. The CHPID type for FICON channels in a point-to-point configuration is required and must be specified as TYPE=FC. for a detailed description of statements and keywords. Refer to IOCP User’s Guide. The switch address is the Domain ID specified in the FICON Director. Note that not all the keywords allowed in the CNTLUNIT statement are described here. The TOK keyword provides information about the source file used to build the IOCDS.ibm. Note that not all keywords allowed in the CHPID statement are described here. TYPE SWITCH CNTLUNIT statement The CNTLUNIT statement contains keywords that define the characteristics of a control unit. SYSTEM The SYSTEM keyword specifies the machine limits and rules that are used for verification of input data set. SB10-7037. PATH LINK The PATH keyword specifies the channel path in each channel subsystem (CSS) attached to the control unit. The device type of the server using the IOCDS has to be specified in the ID keyword. a CSS number can be assigned to the logical partition. The link address can be Appendix C. PART The PART keyword allows you to specify an LPAR name and to assign the MIF ID to a partition.

CUNUMBR The CUNUMBR keyword specifies the control unit number to which the device is attached. IODEVICE statement The IODEVICE statement specifies the characteristics of an I/O device and defines to which CU the devices is attached. Specifying a two-byte link address requires a special feature (High Integrity Fabric) and setup in the FICON Director. contact your System Service Representative supporting the FICON control units. the logical address is specified as two hexadecimal digits in the range 00–FE. 258 FICON Planning and Implementation Guide . To determine a product’s logical addressing information. Not all FICON control units support logical addressing. CUADD For FICON native (FC) channel paths. UNITADD The UNITADD keyword specifies the unit addresses of the I/O devices that the control unit recognizes.specified with one or two bytes.

D Appendix D. switched point-to-point. 2009. © Copyright IBM Corp. Configuring the DS8000 for FICON This appendix describes how to configure the I/O ports on a DS8000 to communicate with a z10 EC server using the native FICON protocol. The configuration steps described are applicable to point-to-point. All rights reserved. 259 . and cascaded FICON Director topologies. 2005. 2006.

xxx. The following steps describe how to set up the ports for FICON use: 1.Getting started The LX ports 0003 and 0242 in our DS8000 are used to connect to either z10 FICON Express8 ports or to FICON Director ports. Click Storage images to continue.xxx. Start a Web browser on a workstation that has network connectivity to the DS8000 HMC.xxx. and then press Enter. The welcome window will appear as shown in Figure D-1. for example http://xxx. type in the user ID and password. 3.xxx is the IP address of the DS8000 HMC).xxx. The Storage images panel is displayed next (see Figure D-2 on page 261). When the logon panel is displayed. 2. 260 FICON Planning and Implementation Guide . Click Storage Images Figure D-1 Welcome to DS8000 4. Type the address of the DS8000 in the Web browser address field.xxx:8451/DS8000/Login (where xxx.

Select Change to FICON Select Port Figure D-3 DS8000 Configure I/O Ports 7. Appendix D. Click the check box of the port number to select the port you want to change. In the Select Action pull-down menu. The Configure I/O Ports panel will be displayed as shown in Figure D-3. On the Storage Images window click the check box on the left of the image name to select the Storage Image. Configuring the DS8000 for FICON 261 . click Configure I/O Ports. click Change to FICON. This will specify the port to use native FICON protocol for communication. Complete these steps for all ports desired for native FICON protocol. In the Select Action pull-down menu. 6. 9.Select Storage Image Click on Configure I/O Ports Figure D-2 DS8000 Storage Images 5. 8.

Finally. Each card can have four ports (T0-T3). The rest of the notation (shown in Table D-1 on page 263) is the same for rack 2 (R2). A rack can have four I/O drawers (I1-I4). check that all ports using native FICON protocol are specified correctly. Rack 1 (R1) is the rack with the controllers and the HMC in it. A DS8000 can have two racks with I/O drawers installed. FICON is selected for port 0003 Figure D-4 DS8000 I/O ports configured for FICON For our configuration we need to have I/O ports 0003 and 0242 set up for FICON in the controller. DS8000 port layout The DS8000 has two port layouts: a physical port layout and a logical port layout. 262 FICON Planning and Implementation Guide . and it consists of six card slots (C1-C6). The physical port layout is written in the following way: R1-I1-C1-T0. See Figure D-4 for an example of the Configure I/O Ports panel (port 0003 is set up and online).10.

Table D-2 Logical port numbering for rack1 (rear view) I1 000 001 002 003 C1 200 201 202 203 C1 010 011 012 013 C2 210 211 212 213 C2 020 021 022 023 C3 220 221 222 223 C3 I3 R I O R I O 030 031 032 033 C4 230 231 232 233 C4 040 041 042 043 C5 240 241 242 243 C5 050 051 052 053 C6 250 251 252 253 C6 100 101 102 103 C1 300 301 302 303 C1 110 111 112 113 C2 310 311 312 313 C2 120 121 122 123 C3 320 321 322 323 C3 I4 R I O R I O I2 130 131 132 133 C4 330 331 332 333 C4 140 141 142 143 C5 340 341 342 342 C5 150 151 152 153 C6 350 351 352 353 C6 Table D-3 on page 264 shows the last three digits of the logical port layout for rack 2. Appendix D. Table D-2 shows the last three digits (I0xxx) of the logical port layout for rack 1. Configuring the DS8000 for FICON 263 .Table D-1 Physical port numbering (rear view) I1 T0 T1 T2 T3 C1 T0 T1 T2 T3 C1 T0 T1 T2 T3 C2 T0 T1 T2 T3 C2 T0 T1 T2 T3 C3 T0 T1 T2 T3 C3 I3 R I O R I O T0 T1 T2 T3 C4 T0 T1 T2 T3 C4 T0 T1 T2 T3 C5 T0 T1 T2 T3 C5 T0 T1 T2 T3 C6 T0 T1 T2 T3 C6 T0 T1 T2 T3 C1 T0 T1 T2 T3 C1 T0 T1 T2 T3 C2 T0 T1 T2 T3 C2 T0 T1 T2 T3 C3 T0 T1 T2 T3 C3 I4 R I O R I O I2 T0 T1 T2 T3 C4 T0 T1 T2 T3 C4 T0 T1 T2 T3 C5 T0 T1 T2 T3 C5 T0 T1 T2 T3 C6 T0 T1 T2 T3 C6 The logical port numbering is written as I0000 (it can also be written without the I in front).

Table D-3 Logical port numbering for rack 2 (rear view) I1 400 401 402 403 C1 600 601 602 603 C1 410 411 412 413 C2 610 611 612 613 C2 420 421 422 423 C3 620 621 622 623 C3 I3 R I O R I O 430 431 432 433 C4 630 631 632 633 C4 440 441 442 443 C5 640 641 642 643 C5 450 451 452 453 C6 650 651 652 653 C6 500 501 502 503 C1 700 701 702 703 C1 510 511 512 513 C2 710 711 712 713 C2 520 521 522 523 C3 720 721 722 723 C3 I4 R I O R I O I2 530 531 532 533 C4 730 731 732 733 C4 540 541 542 543 C5 740 741 742 742 C5 550 551 552 553 C6 750 751 752 753 C6 264 FICON Planning and Implementation Guide .

All rights reserved. Using HMC and SE for problem determination information This appendix explains how you can use the HMC and SE panels to obtain information when performing problem determination or verifying the FICON environment. 2006. 265 . 2005. 2009.E Appendix E. © Copyright IBM Corp.

Users are logged on in SYSPROG mode (User role) using the CLASSIC User Interface (UI) style. but looks quite different. refer to HMC Operations Guide. Double-click CPC. double-click Groups. All panels we discuss here are based on the following: The hardware being used for window captures is a z10 EC Hardware Management Console (HMC) and z10 EC Support Element (SE). 3. For details about this UI style or about how to change the UI style. 3.10.1 which corresponds to Driver level 76D. The panel has four selectable tabs. both running at version 2. However. 4. Double-click Defined CPCs. Right-click CPC and select CPC Details. SC28-6873. On the SE workplace. Logging on to the HMC and SE To log on to the HCM and SE. the information and data you can gather are the same. Log on to the HMC and SE (see “Logging on to the HMC and SE” on page 266 for more information about this topic). 2. Log on at the HMC using SYSPROG (user ID) 2. Double-click Groups. 5. A System z server can also be set to a User Interface style called Tree-Style User Interface (UI). Highlight the CPC that you want to work with. The panels may differ if you are looking at another processor generation or LIC level. Double-click the Single Object Operations (SOO) icon in the CPC Recovery task list. 266 FICON Planning and Implementation Guide . Figure E-1 on page 267 shows the CPC status of a CPC called SCZP101.HMC and SE versions and user IDs The HMC and SE provide System z and FICON channel path information from a hardware point of view. do the following: 1. We discuss the contents of the Instance Information and Product Information tabs. Displaying System z server CPC details The CPC details work area can be opened by performing the following steps: 1. This UI style provides the exact same information. 4.

Figure E-1 CPC Details . the following information can be retrieved (the values shown reflect our scenario): The IOCDS slot used to Activate the CPC: A2 The IOCDS name used during Activation: IODF00 The Activation Profile associated with the CEC: SCZP201 The Activation Profile last used to Activate the CEC: SCZP201 Figure E-2 on page 268 shows CPC product information. Using HMC and SE for problem determination information 267 . Appendix E.Instance Information panel As highlighted in the CPC Instance Information panel.

Displaying the System z server View Frame Layout panel Follow these steps to display the View Frame Layout panel on the SE: 1. double-click Groups. Click the CPC icon and select View Frame Layout from the CPC Configuration task. 3. The Channel to PCHID Assignment panel shows important information about the PCHIDs and CHPIDs and the adaptor types and their physical locations. 4. 2. Double-click CPC. On the SE workplace. 268 FICON Planning and Implementation Guide . We discuss two SE panels: The View Frame Layout panel shows the location of the installed I/O cages where the FICON channels are normally installed.Figure E-2 CPC Details .CPC Product Information As highlighted in the CPC Product Information panel shown in Figure E-2. Log on to the HMC and SE (see “Logging on to the HMC and SE” on page 266). the following information can be retrieved from this panel: The machine type and hardware model: 2097 / E28 The machine serial number: 02-001DE50 The Model-Capacity Identifier: 714 The Model-Temporary Capacity Identifier:714 The Model-Permanent Capacity Identifier: 714 Finding a physical resource on a System z server The following steps explain how to find a physical resource such as a FICON channel in the System z server.

as shown in the details. On the SE workplace. as shown in Figure E-4 on page 270. 3.s Note: The HMC has a similar panel under CPC configuration tasks that allows the Frame Layout panel to be viewed. Using HMC and SE for problem determination information 269 . 4. 2. Figure E-3 SE View Frame Layout panel Displaying the Channel to PCHID Assignment panel Follow these steps to display the Channel to PCHID Assignment panel on the SE: 1. Click the CPC icon and select Channel PCHID Assignment from the CPC Configuration task. Displaying the View Frame Layout panel Figure E-3 shows the View Frame Layout SE panel. Log on to the HMC and SE (see “Logging on to the HMC and SE” on page 266). double-click Groups. The PCHID Assignment panel appears. Double-click CPC. Selecting this option provides additional information about the frames and cages. Clicking over the presented cages causes a “details view” to display. Appendix E.

double-click Images and select the LPAR you are working from. Because you selected CPC. Card Slot and Jack (card port) Book and Fanout Provides information about what fanout is used to access the channel. On the SE workplace. 4. Select the PCHID or CHPID of interest and double-click over the icon to display the PCHID or CHPID details. double-click Groups.Figure E-4 Channel PCHID Assignment In addition to showing search and different view modes.the channel’s corresponding PCHID number – CSS. Double-click CPC if you know the PCHID number associated. 5. 3. its book location and port Channel State – The current state of the channel. this panel displays information about the channels in use such as: Channel Location Physical location of a channel including Cage. 270 FICON Planning and Implementation Guide .CHPID . Right-click the CPC and select Channels. Otherwise. such as Online and Offline – PCHID . the panel displays the CHPIDS associated to the selected partition.the CHPID and CSSs that this channel belongs to – Card Type . this displays the PCHIDs.the channel’s corresponding hardware Displaying individual FICON channel information Follow these steps to display detailed FICON channel information: 1. If you had selected Images. Log on to the HMC and SE (see “Logging on to the HMC and SE” on page 266). 2.

the threshold established limit number is 4. Note that there is a panel called Status in the field. This indicates that the CHPID has been taken offline. This condition is caused by the defined device not matching the attached device. When a value of 0 (zero) is reached. This is a channel status condition that will appear when the IFCC events have exceeded a coded preset value.Figure E-5 shows the PCHID Details panel for PCHID 05A2. this will be indicated in the CHPID status/state. The threshold value is shown in the Analyze Channel Status SE PD panel. The number of frames that make up a Appendix E. Definition Error Note: When a definition error is detected. Figure E-5 PCHID 05A2 Details panel This list shows the possible state or status that a particular PCHID/CHPID might be in: Stopped Operating Stand-by Stand-by/Reserved IFCC Threshold Exceeded This state indicates that the channel is defined. the CHPID icon will default to the icon that represents a FICON Converter CHPID. The channel is a FICON channel but it is not compatible with the channel type defined in the IOCDS. but not in use. Sequence Time Out A FICON “sequence” is a FICON frame or a group of FICON frames transmitted between the Channel and the CU or between the CU and the channel. For FICON channels. Using HMC and SE for problem determination information 271 . For each IFCC detected. This is the normal indication of a working CHPID. and it is not necessarily representative of the defined channel type. this value will be decremented by the code. The status reflects the current state of the PCHID or CHPID. This means that the channel has been put in “service” mode for repair or test.

Displaying the Analyze Channel Information panels Follow these steps to open the Analyze Channel Information (PD) panels: 1. 272 FICON Planning and Implementation Guide . the panel will display the CHPIDS associated to the selected partition. The Analyze Channel Information panel is shown in Figure E-6 on page 273. A Sequence Time Out is detected by the channel or the CU when an ongoing exchange does not detect the arrival of a required sequence. this displays the PCHIDs. Select Analyze Channel Information and click OK. Double-click Channel Problem Determination located in the CHPID Operations task list. An exchange represents an I/O operation in the FICON implementation. double-click Images and select the LPAR you are working from. Double-click CPC if you know the PCHID number associated. Because you selected CPC. On the SE workplace.sequence is determined by specific fields inside the frames themselves. 4.” It reports that the FICON channel-attached control unit or FICON Director port does not support the presented FICON service parameters. If you had selected Image. 5. Sequence Not Permitted This error is reported in the Analyze Channel information panel as “Illegal Sequence. Otherwise. 2. 7. Sequences are logically grouped to form an exchange. double-click Groups. Log on to the HMC and SE (see “Logging on to the HMC and SE” on page 266). Note: The SE will post a panel for an Image query if you select a PCHID instead of a CHPID. 6. Highlight the channel that you want to view. Right-click the CPC and select Channels. 3.

The IFCC threshold is has an initial value of 4 and decrements at every occurrence of an IFCC on this channel. when not equal to 0 (zero). the entry port and the destination port belong to the same switch number). this is an indication that cascaded switch topology is being used. The Error code. also note that CHPID 7E is reported as Spanned. Using HMC and SE for problem determination information 273 . The block labeled Error Status shows the link status. notice that Partition ID 23 is the LPAR ID associated to the partition where this panel was selected from. meaning that it has been shared between LPARs in different CSSs and is associated to CHPID 5A2.7E Note the blocks that are highlighted: On the top left side. This block also displays the Error code and the Bit Error Counter (Ber) accumulated values for the CHPID. The block on the top right side labeled Hardware Card type Cascade definition indicates that this channel was defined using 2 byte control unit link address. The Temporary error threshold works the same way as the IFCC threshold. (In this case. in the block labeled Definition Information. It also reports that CHPID 7E is connected to Switch number 65. When the value reaches 0 (zero). the indication of a IFCC Threshold Exceeded will be associated to the CHPID icon on the SE desktop.Figure E-6 Analyze Channel Information PD panel for CHPID 2. Note: HCD allows a dual-byte address definition even when a single FICON Director is used in the configuration. the connection port. Normally. The SAP affinity indicates which of the SAPs this CHPID is associated to. The block labeled Status Information indicates the status and the state information of this CHPID. and the SAP affinity associated to the CHPID. MIF image ID 3 corresponds to the LPAR MIF ID on CSS 2. Appendix E. may be displayed using a panel button called Error details.

This field represents the serial number of the node. The System z server adds a new flavor to the Tag field. Invalid means there has not been any port login. It has a direct relationship with the information provided about the Hardware type and Subtype presented in the top right corner of the panel. Node status Valid means there has been a successful port login.CHPID for System z server. Attached indicates the shown Node Descriptor (ND) information is from the attached node. as explained here: . and CSS2 represents CHPID 7ETable E-1 on page 275 lists the CSS-to-tag relationship. Table E-1 on page 275 illustrates the CSS/Tag relationship. and you must also refer to the vendor’s documentation to understand how to decode this field. For example. This field indicates the plant where the node was manufactured. The tag value in the top line gives the corresponding CSS IDs that the node has been defined to. Understanding Node (ND) information displayed on this panel The data listed here is received from the request node identification (RNID) during the initialization of the link: Node type Self means the shown Node Descriptor (ND) information is from the channel itself. with Tag = A07E. the A indicates that the node is defined to CSS0. VALID BUT NOT CURRENT means there has been a port login. The Tag field represents the CSS.The block labeled Card Type Connection rate shows the FICON card description type associated with this CHPID. For the attached node. . but the link got dropped afterwards. you need to know whether the tag is vendor-dependent. 274 FICON Planning and Implementation Guide . Type/Model Plant Sequence number Tag This field indicates the machine type and model of the node.The first character of the Tag field represents the CSS.The last two digits of the Tag field represent the CHPID (the second digit showing a zero (0) is reserved).

Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. In our example. On the SE workplace. This is a unique 64-bit address that is used to identify the port in an FC topology. 4. 7. Select the option Search by Device Number (you can also search by a unit address or subchannel number. Log on to the HMC and SE (see “Logging on to the HMC and SE” on page 266). the PCHIDs will display. Otherwise. Appendix E. If you select CPC. 3. Right-click the CPC or Images and select Channels.Table E-1 CSS-to-tag relationship CSS\Tag 0 1 2 3 X X X X X X X X X X X X X X X X X 1 2 3 4 5 6 7 8 X 9 X A X B X C X X D X X E X X X F X X X X World Wide Node Name World Wide Port Name This is a unique 64-bit address that is used to identify the node in an FC topology. 9. 2. Using HMC and SE for problem determination information 275 . Double-click Channel Problem Determination. if appropriate) and click OK. Enter the device number (or unit address or subchannel number. 6. This address is assigned by the manufacturer. Highlight the channel you that want to view. 5. which is located in the CHPID Operations task list. This address is assigned by the manufacturer. Displaying the Analyze Subchannel Information PD panel Follow these steps to open the Analyze Subchannel Information panel: 1. double-click Groups. If you select Images. double-click Images and select the LPAR you are working from. 8. Double-click CPC if you know the PCHID number associated. the CHPIDs associated to a logical partition will display. we used device number D000 (see Figure E-7 on page 276). Select the option Analyze Subchannel Information and click OK. depending on what you selected in the previous step) and select OK.

Logical Path Established: this field indicates the paths to this device that successfully went through the pathing process.Associated CHPID number.Logical Path Mask: the bits show which CHPIDS can be used to access the device. 276 FICON Planning and Implementation Guide .PCHID. During IPL device mapping process. Bit 1.CHPID1. The following fields are shown: CHPID . Each bit position represents one channel path. Bit 0. The list of candidate CHPIDs is shown in the top right corner. corresponds to CSS. or the first position bit from the left. PNOM . LPE .PCHI.CHPID0. The Pathing information is displayed in the block in the center of the panel. corresponds to CSS. This Pathing information has various fields that are bit-position-relative to the defined channel paths to this device.Figure E-7 Analyze Subchannel Information PD panel Note the fields that are highlighted: The field Irpt parm: 02362250 displays this subchannel’s associated UCB SCP absolute address. or the second position bit from the left. LPM . and so on. the subchannel is connected to its UCW and this field contains the respective pointer to the UCB in SCP memory.Path Non-Operational Mask: this field indicates which paths from the list of candidate CHPIDs are in the Not Operational state.

PAM . Double-click Channel Problem Determination. Highlight the channel you want to view. 3. depending on what you have selected in the previous step) and select OK. OFFLINE z/OS command is issued from the operating system console. Using HMC and SE for problem determination information 277 . the PCHIDs will display. The Control Unit Header PD panel is shown in Figure E-8 on page 278. 5. – The CF CHP(xx). On the SE workplace. 9. 2.Path Available Mask: this field indicates the Physically Available channel paths. Initially (at System Reset).Path Installed Mask: this field indicates the channel paths that are defined in the IOCDS. the CHPIDs associated to a logical partition will display. Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. It is updated when the Ending Status is received from the device. 8. 4. PIM .Last Path Used Mask: this field indicates the last path used to access the device when successful or any abnormal condition occurred for the I/O operation. 6. 7. double-click Images and select the LPAR you are working from. Log on to the HMC and SE (see “Logging on to the HMC and SE” on page 266). Displaying the Analyze Control Unit Header PD panel Follow these steps to display the Control Unit Header Problem Determination panel: 1. The initial value will be the paths defined in the IOCDS for the LPAR. Otherwise. Select the option Analyze Control Unit Header and click OK.Path Operational Mask: this field contains the paths online to the device.LPUM . which is located in the CHPID Operations task list. this value is set to FF (all ones) and the corresponding bit will be turned OFF when the first attempt to access the device fails (Inoperative). This bit will be turned off when: – The corresponding CHPID is configured OFF from the HMC. double-click Groups. Enter the device number (or unit address or subchannel number. If you select CPC. POM . Select the option Search by Device Number (you can also search by a unit address or subchannel number if appropriate) and click OK. Appendix E. Right-click the CPC or Image and select Channels. Double-click CPC if you know the PCHID number associated. If you select Images.

along with the CHPIDs associated to these paths and their respective FICON Director links. Log on to the HMC and SE (see “Logging on to the HMC and SE” on page 266). Refer to “DS8000 port layout” on page 262 for the logical port layout of a 2107. The output presented shows: This indicates all possible paths to get to the entered device. Plant. Number and Tag All together. Path CHPIDS Linkqq All the possible paths to get to the entered device are displayed. 278 FICON Planning and Implementation Guide . Type/Model. This indicates the destination link address in use by each of the CHPIDs and respective paths. Using Table D-2 on page 263 to translate tag value 0003 to a physical port results in the following: Rack 1 (R1) I/O enclosure 1 (I1) Card 1 (C1) Port 3 (T3) Displaying the Analyze Paths to a Device PD panel Follow these steps to display the Analyze Paths to a Device panel: 1. MFG. these represent the Control Unit RNID (Remote Node Identification) information. This indicates the channels associated to the displayed paths.l Figure E-8 Analyze Control Unit Header PD panel The starting point to obtain this data is provided by the device number entered when selecting this PD action. which is connected to a 2107 using tag 0003. Seq. In our example we highlighted the FICON CPID 2. where applicable.7E in the control unit header frame. Node Status This indicates the current status of the node. along with the tags (which represent the attaching Host Bay Adapters) that the channels connect to. This device’s control unit characteristics are also displayed. Node Type This indicates the type of attachment to the CHPIDs and Paths.

2. Figure E-9 Analyze Paths to a Device PD panel It can display important device path information. the PCHIDs will display. 9. The Analyze Path to a Device panel is shown in Figure E-9. On the SE workplace. as follows: The device’s associated subchannel. Double-click CPC if you know the PCHID number associated. which is located in the CHPID Operations task list. This shows if there is a FICON Director in the link. Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. double-click Images and select the LPAR you are working from. double-click Groups. Otherwise. 3. 8. the CHPIDs associated to a logical partition will display. Enter the device number (or unit address or subchannel number. If you select Images. 5. Select the option Search by Device Number (you can also search by a unit address or subchannel number if appropriate) and click OK. This is the associated CHPID number. depending on what you have selected in the previous step) and select OK. Right-click the CPC or Image and select Channels. with the following additional information: Avail CHPID Switch This shows if that path is available at that instance. Double-click Channel Problem Determination. Highlight the channel you want to view. Select the option Analyze Paths to a Device and click OK. If you select CPC. Using HMC and SE for problem determination information 279 . All paths by which the device can be accessed (this is relevant only for the partition the channel path has been selected from). Appendix E. 6. 7. 4.

Displaying the Analyze Device Status PD panel Follow these steps to open the Analyze Paths to a Device panel: 1. 280 FICON Planning and Implementation Guide . Otherwise. 9. Enter the device number (or unit address or subchannel number. which is located in the CHPID Operations task list. Double-click Channel Problem Determination. Highlight the channel you want to view. Select the option Analyze Device Status and click OK. double-click Images and select the LPAR you are working from. On the SE workplace. Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. This is the destination link address (the port address the CU is connected to). Select the option Search by Device Number (you can also search by a unit address or subchannel number if appropriate) and click OK. 2. Right-click the CPC or Image and select Channels. 3. double-click Groups. This is the logical CU that the device is connected to.Switch Number Linkaddr Cuadd This is the FICON Director number (if applicable). 6. The Analyze Device Status PD panel is shown in Figure E-10 on page 281. 4. 7. the CHPIDs associated to a logical partition will display. Log on to the HMC and SE (see “Logging on to the HMC and SE” on page 266). depending on what you selected in the previous step) and select OK. 5. If you select CPC. the PCHIDs will display. If you select Images. 8. Double-click CPC if you know the PCHID number associated.

which is located in the CHPID Operations task list Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. The Analyze Serial Link Status panel is shown in Figure E-11 on page 282. On the SE workplace. Using HMC and SE for problem determination information 281 . double-click Groups. Highlight the channel you want to view. Displaying the Analyze Serial Link Status PD panel Follow these steps to open the Analyze Paths to a Device panel: 1. Note that a Filter Status button is provided to find a specific status a particular device. Otherwise. 3. 5. 2. double-click Images and select the LPAR you are working from. Appendix E. If you select Images.Figure E-10 Analyze Device Status panel The Analyze Device Status panel reports the current state of all devices pertaining to a Logical Control Unit. the PCHIDs will display. Select the option Analyze Serial Link Status and click OK. Double-click CPC if you know the PCHID number associated. If you select CPC. the CHPIDs associated to a logical partition will display. Double-click Channel Problem Determination. Log on to the HMC and SE (see “Logging on to the HMC and SE” on page 266). 7. Right-click the CPC or Images and select Channels. 4. 6.

A series of logical Control Units (residing in the same physical CU) being accessed from Channel 2. However. Channel link address . Switch number . as well. The status of all logical link addresses can be seen in this panel. Note. CHPID .This is the logical channel consisting of CSS and CHPID.This is the MIF ID of the image we selected the channel from.This is the switch number of the FICON Director that the channel is connected to. this would show a status of INITIALIZATION PENDING.Figure E-11 Analyze Serial Link Status PD panel The following logical pathing information is displayed: MIF Image ID . that a status of INITIALIZATION PENDING is not necessarily an error indication.This is the total number of switches in the links. This provides information about the logical CU initialization from a channel point of view. Any status other than INITIALIZATION COMPLETE indicates a potential problem that may need to be investigated. however.7E show the status INITIALIZATION COMPLETE. For example. if a channel is trying to establish a FCTC connection to another server or image but the LPAR is not activated. Switch number valid .7E from MIF Image 3 connected to Switch number 65 port address 03. It depends on whether the logical CU is expected to be available. 282 FICON Planning and Implementation Guide .This is the channel type as defined in the IOCP. Channel type . The given CU address cannot be accessed if the status is other than INTIALIZATION COMPLETE. Channel 2.This is the link address (port address) that the channel is connected to. the status is expected to change after the target LPAR is activated.

Figure E-12 New PLOGI Initialization message Appendix E.Init Failure .Link Reject .This indicates that there is a reset initiated by the CU itself in progress.Link Reject .Link Reject . in plain English.Channel Detected Error Internal CTC Initialization Complete Link Level Error . CU reset in progress . Channel/CU mismatch . Using HMC and SE for problem determination information 283 .FICON PLOGI ELS Error Link Level Error .CU Busy Initialization Complete .FICON FLOGI ELS Error Link Level Error .This indicates that the logical CU address (defined in the CU) does not match its definition in the IOC.Undefined Destination Address Switch Port Malfunction .This indicates that the CU ran out of logical CUs.Port Destination Malfunction Switch Port Not Available .Link Reject .FICON LIR ELS Error Link Level Error .Port Reject .– Other typical error messages are: • • • • • • • • • • • • • • • • • • • • • • • • • • • • Initialization Complete . Figure E-12 shows one of the new messages issued when initialization failures occur.Channel/CU mismatch CU Resources Exceeded .P PLOGI error .No Resources Available Link Level Error .FICON RNID ELS Error Link Level Error . With this new implementation the panel will show.Dynamic Switch Port Busy Initialization in Progress CU Reset in Progress Remove Logical Path in Progress Link Configuration Problem .Acquire Address Error Link Configuration Problem .This indicates that the link initialization did not succeed to the Destination Port. what went wrong during the CU link initialization.Initialization Failure . Note: The latest implementation of the System z Server SE Problem Determination facility improves the PLOGI and FLOGI error messages substantially.Port Reject .Transmission Error Link Level Error .Destination Address not Valid Link Level Error .Port Reject .FICON Invalid Attachment CU resources exceeded .Port Reject .Uninstalled Link Control Function Link Level Error .Link Reject .Address not Valid Switch Port Not Defined .Unrecognized Device Level Link Configuration Problem .FICON SCR ELS Error Link Level Error .Port Intervention Required Link Configuration Problem .

the CHPIDs associated to a logical partition display. which is located in the CHPID Operations task list. Highlight the channel you want to view. 4. Select the option Analyze Link Error Statistics Block and click OK. double-click Groups. If you select Images. Figure E-13 Link Error Statistics Block PD panel The Analyze Link Error Statistics Block (LESB) is a new implementation on the System z Server’s Problem Determination SE panels. If you select CPC. The Analyze Link Error Statistics Block PD panel is shown in Figure E-13. the PCHIDs display. double-click Images and select the LPAR you are working from. 6. Otherwise. 5. Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. Log on to the HMC and SE (see “Logging on to the HMC and SE” on page 266). 284 FICON Planning and Implementation Guide . It displays the Channel N-Port LESB concurrently with channel operations. Double-click CPC if you know the PCHID number associated. 3. On the SE workplace. Right-click the CPC or Images and select Channels.Displaying the Analyze Link Error Statistics Block PD panel Follow these steps to open the Analyze Paths to a Device panel: 1. 2. 7. Double-click Channel Problem Determination.

A refresh provides a dynamic display of the counters. Using HMC and SE for problem determination information 285 .5. “FICON Purge Path Extended” on page 222. refer to 10. The information gathered from the Channel N-Port is also recorded in the EREP SLH records when a Purge Path Extended (PPE) operation is performed.The various types of errors are accumulated to provide you with a view of how the channel is operating. For more information about the PPE process. Appendix E.

286 FICON Planning and Implementation Guide .

2009.F Appendix F. 2005. © Copyright IBM Corp. 2006. Useful z/OS commands This appendix discusses the information pertaining to the FICON environment that can be obtained from various z/OS commands. 287 . All rights reserved.

as well as its MIF and LPAR IDs.The output shows the CPU type and serial number. the CPC name. as well as the CPU Node Descriptor (ND). the System Information taken from the Store System Information (STSI) instruction. in this case). Example: F-1 Displaying the CPU characteristics D M=CPU IEE174I 11.13 DISPLAY M 944 PROCESSOR STATUS ID CPU SERIAL 00 + 23DE502097 01 + 23DE502097 02 03 04 -A 05 -A 06 +I 23DE502097 07 -I 288 FICON Planning and Implementation Guide .38.Using z/OS commands for problem determination Figure F-1 represents the FICON environment that we used to produce the output of the z/OS commands discussed in this appendix. the CSS ID where this partition is running. zIIPs and zAAPs configured. The output shows the General CPs. z10 EC SCZP201 LPAR SC30 A23 z/OS V1R10 FICON Express8 LX CSS2 MIF 3 PCHID 5A2 CHPID 7E System Storage SAN384B Switch # 66 Switch @ 66 Port 8D Port C4 Port 88 Port CF DS8000 CU# D000 LX 0003 D0xx ISL Port 03 Port 2D Port 34 LPAR SC31 A24 z/OS V1R10 FICON Express8 LX CSS2 MIF 4 PCHID 5D2 CHPID 82 Switch # 65 Switch @ 65 LX 0242 D1xx CU# D100 Port 1B * All cable connectors are LC Duplex type SAN768B Figure F-1 Our FICON configuration Displaying system status using D M=CPU Example F-1 shows the output of a z/OS D M=CPU command. It also defines the CPU characteristics associated to a specific partition (LPAR A23.

ID to be specified for the CPC.02.E26. HCD/HCM also requires a Proc. Useful z/OS commands 289 . or deleting LPARs requires a new IOCDS and a Power-on Reset (POR) on System z9. DOES NOT EXIST W WLM-MANAGED N NOT AVAILABLE A APPLICATION ASSIST PROCESSOR (zAAP) I INTEGRATED INFORMATION PROCESSOR (zIIP) CPC ND CENTRAL PROCESSING COMPLEX NODE DESCRIPTOR CPC SI SYSTEM INFORMATION FROM STSI INSTRUCTION CPC ID CENTRAL PROCESSING COMPLEX IDENTIFIER CPC NAME CENTRAL PROCESSING COMPLEX NAME LP NAME LOGICAL PARTITION NAME LP ID LOGICAL PARTITION IDENTIFIER CSS ID CHANNEL SUBSYSTEM IDENTIFIER MIF ID MULTIPLE IMAGE FACILITY IMAGE IDENTIFIER The other fields in the display output have the following meanings: CPC name This reflects the CPC object name customized to the HMC at System z installation time. CSS ID This indicates in which CSS this LPAR has been created.02. The Enterprise Class (EC) models support four CSSs in total. Each CSS supports two subchannel sets: SS0 and SS1. between 1 and 15.714. channel subsystems. logical processors and cryptographic co-processors. that is associated to a partition in one CSS. The name has to be unique for each logical partition across all four Channel Subsystems (CSSs) on System z. subchannel sets. LP name This is defined in the HCD/HCM logical partition definition or in the IOCP RESOURCE statement.IBM. The number of CSSs that a System z supports depends on the machine model. Increasing or decreasing the number of CSSs in use in a z9 requires a POR and is concurrent on z10. The CPC name is also defined in the HCD/HCM (a CPC can only be defined once). Changing. and is fully concurrent on System z10. Note: System z10 (2097 and 2098) implemented a fixed Hardware System Area (HSA) storage approach that allows for concurrent adding logical partitions.OFFLINE . The Business Class (BC) models for the z9 and z10 support two CSS. MIF ID This is a value.000000000001DE50 Model: E26 CPC ID = 00 CPC NAME = CZP201 LP NAME = A23 LP ID = 23 CSS ID = 2 MIF ID = 3 + ONLINE .CPC ND = 002097. The MIF ID is used by the System z channel subsystem and channels to identify the source of an I/O Appendix F.00000001DE50 CPC SI = 2097.10 or later. The MIF ID must be unique for each logical partition in a System z CSS. adding. Adding logical processors to a LPAR requires z/OS 1.IBM. The MIF ID numbers are defined in the HCD/HCM in the IOCP RESOURCE statement.

10.00 LICENSE = z/OS USED LOAD01 IN SYS0. The LP ID number.CONFIG Example F-3 on page 291 displays the following information about the I/O configuration in use by the System z: The IODF data set name (ACTIVE IODF DATA SET = SYS6.16.IPLPARM ON C730 ARCHLVL = 2 MTLSHARE = N IEASYM LIST = XX IEASYS LIST = (00) (OP) IODF DEVICE C730 IPL DEVICE DC0D VOLUME Z1ARC1 The output displays the following information: The z/OS release (RELEASE z/OS 01.request.00)that this partition is running The time stamp.10 ON 03/12/2009 RELEASE z/OS 01.IODF75) The active CSS and subchannel sets in use (ACTIVE CSS:2 SUBCHANNEL SETS CONFIGURED:0. LP ID The logical partition ID (LP ID) is specified in the Logical Partition Image Profile using the HMC. 1) The token that will allow dynamic changes to the configuration A match between the TOKEN IODF and the ACTIVE IODF DATA SET names will indicate whether the system allows dynamic hardware and software changes to the I/O configuration. Example: F-2 Displaying the z/OS IPL information D IPLINFO IEE254I 09.IPLPARM ON C730) The system architecture level (ARCHLVL = 2). Displaying the I/O configuration using D IOS. may lead to I/O-related and configuration-related problems.58. from x’00’ to x’3F. or selecting an unappropriated IODF device.10 ON 03/12/2009) The LOAD member selected from the PARMLIB (USED LOAD01 IN SYS0.16. Displaying additional z/OS information using D IPLINFO Example F-2 displays the output of a z/OS D IPLINFO command.10. which shows that this partition was IPLed (SYSTEM IPLED AT 09. 290 FICON Planning and Implementation Guide .13 IPLINFO DISPLAY 899 SYSTEM IPLED AT 09. indicating a 64-bit operating system in this case The IODF device The IPL device and its VOLSER (IPL DEVICE DC0D VOLUME Z1ARC1) Note: IPLing the system from the wrong device. must be unique across all partitions on a System z. You need to know the MIF ID to resolve LP-related problems or to resolve failures in establishing a FICON logical path.

LOGICAL CONTROL UNITS 4050 SS 0 SUBCHANNELS 56520 SS 1 SUBCHANNELS 65535 The output displays: The total number of physical control units defined in the configuration: 8064 The number of logical control units per Channel Subsystem The number of subchannels defined in each CSS’s subchannel set (0 and 1) Appendix F. 1 CHANNEL MEASUREMENT BLOCK FACILITY IS ACTIVE Displaying HSA usage using D IOS. Note: The maximum number of subchannels on subchannel set 0 (SS0) is 65280.IODF75 CONFIGURATION ID = TEST2094 EDT ID = 01 TOKEN: PROCESSOR DATE TIME DESCRIPTION SOURCE: SCZP201 09-05-04 09:50:17 SYS6 IODF75 ACTIVE CSS: 2 SUBCHANNEL SETS CONFIGURED: 0.00.1.LOGICAL CONTROL UNITS 3980 SS 0 SUBCHANNELS 51215 SS 1 SUBCHANNELS 65535 CSS 1 . which contains hardware controls and definition information about a particular device. Useful z/OS commands 291 .LOGICAL CONTROL UNITS 3987 SS 0 SUBCHANNELS 51375 SS 1 SUBCHANNELS 65535 CSS 2 .22 I/O CONFIG DATA 903 HARDWARE SYSTEM AREA AVAILABLE FOR CONFIGURATION CHANGES PHYSICAL CONTROL UNITS 8064 CSS 0 . System z9 BC and z10 BC can have up to two Channel Subsystems (0 and 1) with two subchannel sets each.19 I/O CONFIG DATA 901 ACTIVE IODF DATA SET = SYS6.01. The maximum number of subchannels on subchannel set 1 (SS1) is 65535.CONFIG(HSA) IOS506I 10.LOGICAL CONTROL UNITS 4004 SS 0 SUBCHANNELS 52515 SS 1 SUBCHANNELS 65535 CSS 3 .CONFIG(HSA)’ z/OS command output D IOS.CONFIG IOS506I 10. 2 and 3) with two subchannel sets each. Note the following points: System z9 EC and z10 EC can have up to four Channel Subsystems (0.Example: F-3 ‘D IOS.CONFIG’ z/OS command output D IOS.CONFIG(HSA) Example F-4 displays information about the amount of physical control units and subchannels in use per subchannel set in each defined channel subsystem (CSS). A subchannel corresponds to a device Unit Control Word (UCW). Example: F-4 D IOS.

where the z/OS Display Units command will not display the status of a defined device. or The next device of that status.dddd. If the z/OS status for the requested device number is displayed.n or D U.1 The Display Units command is useful for determining the z/OS status of a device.. The example shows the normal result of displaying the z/OS status of a group of devices. (DASD ALIAS devices is one of these exceptions.Display units command D U Use the z/OS Display Units command as the starting point for determining the status of a device defined to z/OS. and so on. The output from this command is shown in Example F-5 on page 293. for a display units type request. Online and in use by a job or task. If the devices are not defined to z/OS. Determine the defined device type: 3390.n” on page 300. or If a combination of both type and status is used. it is defined to z/OS. Indicate the z/OS status of the device. the statuses include: O F-NRD Allocated A A-BSY Online.dddd.n) is entered. D U.. 292 FICON Planning and Implementation Guide . There are a few exceptions to this.. or that it includes your requested device number. always verify that the device (or devices) displayed in response to a z/OS Display Units command request is for your display unit’s requested device number. if it is not equal. that is. D001.DS P.ALLOC. VOLSER = NE800 Whenever a z/OS Display Units command (or any other general form of this command such as D U. it indicates which volume is currently mounted on that device. Allocated and busy.) The Display Units z/OS command can be used to: Determine if the device is defined to z/OS (exception devices are not displayed). then the device will not be brought online to z/OS.DASD. For devices that support “volumes”. for more information refer to “Using a D M=DEV(dddd) command for an ALIAS device number” on page 299 and to “DEVSERV command . Identify whether the defined z/OS type is not equal to the device being addressed (when trying to bring the device online). for a display units by status. and the z/OS status of the defined paths to the device. meaning that there is an I/O operation in progress for the z/OS device in the channel subsystem image (in the channel subsystem for the logical partition that z/OS is running in).. the devices are not in use by a job or application. z/OS will display the status of the devices that the z/OS SCP has recorded at that time. Offline.dddd. then z/OS will display one of the following pieces of information: The next device number or the next device number of that type. then it will be the next device of that type and status (refer to Example F-6 on page 294) Because of this. 3590. or if they are exception devices. and D002 are online and not allocated. CTC. Devices D000.dddd.

such as JES3 or a non-IBM tape management subsystem. P Reserve Pending PND Offline pending PO Offline pending.4 IEE457I 20.Not Ready O .. F Offline. F indicates that more than one bit is set in the UCB.D000. M A device managed by a device manager. The device has been BOXED (refer to Boxed status notes: below). Enter the display allocation command DU. Example: F-5 z/OS display device status of a group of four devices defined and online D U. SYS Allocated to system.ONLINE OFFLINE This is used when the only status value that needs to be displayed is OFFLINE.52 UNIT STATUS 727 UNIT TYPE STATUS VOLSER VOLSTATE D000 3390 O Z1BRZ1 PRIV/RSDNT D001 3390 O Z18DL2 PRIV/RSDNT D002 3390 O Z18RB1 PRIV/RSDNT D003 3390 A TOTD94 PRIV/RSDNT There are also a few other device statuses that may be returned for a D U command: Automatically switchable. BSY Busy. S .8003. L The release on a device is pending and reserve may or may not have occurred.Mount Pending NRD . Being allocated means the device is in use by a job or an application. RAL Restricted to Allocation. (a paging volume). and also not ready. This is used when a combination of offline and some other status value needs to be displayed (for example. The channel program is temporarily suspended while the system is using the device. R Reserved.UNAVAIL operator command.SYSRES SPD Suspended.Device D003 is both online and allocated. This status value is displayed only if NRD is also displayed on the status line. MTP . Hardware failure. PUL Unload pending. F-NRD). C Console. you can use the z/OS display allocation command to determine who the device is allocated to.1 and it will show who the device is allocated to. If the device is online and allocated. UNAVL The device has been marked as unavailable for allocation by the VARY xxxx.ALLOC. Useful z/OS commands 293 . AS BOX Appendix F. shared DASD or exclusively-assigned device..35..

.UNCOND. the device will come online.dddd.58 UNITS ALLOCATED 722 UNIT JOBNAME ASID JOBNAME ASID JOBNAME ASID JOBNAME ASID D003 DUMPSRV 0005 MVSRECVA 0028 The information displayed shows that the jobs using the device are: Jobname . Example F-7 shows the device number D003 as online.ddd. 294 FICON Planning and Implementation Guide .DUMPSRV Jobname . however.. allocated to two jobs.D003. If the error condition still exists. However.MVSRECVA V dddd. however. the action of varying this device offline will be in a pending status.21 UNIT STATUS 822 UNIT TYPE STATUS VOLSER VOLSTATE 8100 3390 O NW8100 PRIV/RSDNT The D U request shown in this example was for device 8045. D U... device D003 is allocated and in use.ONLINE. This will enable the UCW and perform the online processing to the device. the V dddd. then z/OS will not be able to display the status for that device. Displaying a device not in current configuration If the z/OS display units request is for a device that is not defined in the current OSCONFIG member (in use by z/OS).1 IEE106I 08.ALLOC.8045. and currently in use..ALLOC. Therefore.1 IEE457I 08.offline In Example F-8 on page 295..Boxed status notes: If the reported status of a device is O-BOX (Online and Boxed).n The D U. or if the device is an exception device. Example: F-6 D U request of a non-existing device D U. that status can be cleared by using a vary online with unconditional parameter: V dddd. For this reason.ALLOC. or the requested device by status (when using the display unit status command).. the device can be brought back online with the VARY dddd. notice that the actual display reports the status for device 8100. always check that the devices displayed include your requested device. Its main purpose is to display the allocations of a device when its status has been determined to be Allocated. see in Example F-6.OFFLINE command is issued against device D003.n command is one of the various formats for the Display Units z/OS device commands.07. as shown in Example F-7. Example: F-7 z/OS display device allocation D U.57.ONLINE command. If the reported status is F-BOX (Offline and Boxed). Assuming that the error condition has been resolved. the device may remain in the boxed state.. or the next device number of the requested device by type (when using the display units type command). z/OS will display the next displayable device number.

D U. A group of devices can be varied online using dddd.. that is. It was defined to the operating system as being offline at system IPL time.the CU logical image address. All the operational paths were logically varied offline. CHPID. Link destination link address (destination port address). The device is in an error state.D000. a reserve or assign was lost) and the device was forced offline (it may have been boxed during the recovery process.OFFLINE IEF524I 8003. V dddd. as well).D003.. the CUADD is unique for an LSS. Useful z/OS commands 295 . The definition to access the control unit device is not correct..ONLINE Use the z/OS command shown in Example F-10 on page 296 to bring a single device online. the CUADD is the MIF ID of the target LPAR. There is no operational path to the control-unit or device. Link destination link address (destination port address)..27. For a CTC.1 IEE457I 18.. There are variations to the V dddd. VOLUME TOTD94 PENDING OFFLINE .4 IEE457I 08.12 UNIT STATUS 674 UNIT TYPE STATUS VOLSER D003 3390 OFFLINE VOLSTATE /RSDNT A device can be offline for a number of reasons: It has been varied offline by the operator (this is the case shown in Example F-90. CUADD . that is.dddd or dddd-dddd as variables for the vary command.online command that allow more than one device to be brought online simultaneously. The device is defined to the operating system but not to the channel subsystem. Appendix F. The defined device type and the physical device type (the actual device) do not match.Example: F-8 Varying an allocated device offline V D003. Example: F-9 Device is now offline after the allocation ended D U. An error condition occurred during the use of the device (that is.. For an IBM 2107. The device is defined to the operating system but not to the logical partition that the z/OS is running in. there is no subchannel. The device addressing definition is not correct.. The device is not ready.21 UNIT STATUS 756 UNIT TYPE STATUS VOLSER VOLSTATE D000 3390 O Z1BRZ1 PRIV/RSDNT D001 3390 O Z18DL2 PRIV/RSDNT D002 3390 O Z18RB1 PRIV/RSDNT D003 3390 A-PND TOTD94 PRIV/RSDNT Device in OFFLINE status Example F-9 shows a display reporting that device D003 is in the OFFLINE state. there is no subchannel.16.

900.IBM.ONLINE IEE302I D003 ONLINE A failure to bring a device online can be caused by: There is no UCB for the target device.Example: F-10 V dddd.0000 SCP DEVICE NED = 002107.58 DISPLAY M 907 DEVICE D003 STATUS=ONLINE CHP 7E ENTRY LINK ADDRESS 6503 DEST LINK ADDRESS 6688 PATH ONLINE Y CHP PHYSICALLY ONLINE Y 296 FICON Planning and Implementation Guide . D M=DEV(dddd) The D M=DEV(dddd) command displays the device status for device dddd. Example F-11 shows the output of this display device matrix command.75. ONLINE results V D003.50. Example: F-11 z/OS display output for D M=DEV(dddd) D M=DEV(D003) IEE174I 09.900.00.75.0000000BALB1. None of the defined paths are operational. All paths to the device are offline. in the example). It shows the path status of all paths that have been defined to this single device address (device D003.0300 SCP TOKEN NED = 002107.(cc)) D M=DEV(D003.75.0003 HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW.(7E)) IEE174I 18. Example: F-12 Variation of D M=DEV(dddd.IBM. There is a mismatch of device type between the definition and the actual device. ZHPF Example F-12 is a variation of the D M=DEV(dddd) command where a unique CHPID that accesses the device dddd is selected for display.900.0000000BALB1.IBM.0000000BALB1. The CU and or the device addressing is not correct. The destination link address as well as the entry link address are individually provided for each channel accessing the device. There is no UCW (subchannel) for the target device.12 DISPLAY M 767 DEVICE D003 STATUS=ONLINE CHP 50 54 58 5C 7E 82 ENTRY LINK ADDRESS 0F 0F 1B 1B 6503 651B DEST LINK ADDRESS 2C 2C 0A 0A 6688 66CF PATH ONLINE N N Y Y Y Y CHP PHYSICALLY ONLINE N N Y Y Y Y PATH OPERATIONAL N N Y Y Y Y MANAGED N N N N N N CU NUMBER D000 D000 D000 D000 D000 D000 MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107.

900.75. the UCBVALPH bit is set to On to indicate that the paths to the device were not validated. Port 03. MANAGED = N This field indicates whether the channel is using the Dynamic Channel Management (DCM) facility of z/OS.0000000BALB1. Note that the FICON Director addresses are different for the entry and exit (Dest) Ports. PATH OPERATIONAL = Y This means that the PATH OPERATIONAL field status is changed as a result of the channel subsystem attempting an I/O operation on a path and the path responding with a not operational I/O interface sequence.0003 HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW.OFFLINE command. The CHP PHYSICALLY ONLINE field indicates whether the path is physically available. Validation will occur when the device is varied online.FICON Director @ 65.0000 SCP DEVICE NED = 002107. the following important information is displayed: The device current status: DEVICE D003 STATUS=ONLINE. Information for this field is obtained from the Path Available Mask (PAM) in the Unit Control Word (UCW) or subchannel. If a device is not online when the IPL device pathing process occurs. This field will change as a result of the z/OS CF CHP command. This is an indication that a FICON Director cascading topology is being used.900.900.0000000BALB1. PATH NOT VALIDATED = Y The PATH NOT VALIDATED field may be shown if the device was not online at IPL time. Note: The CHP PHYSICALLY ONLINE field might not be accurate when the CHPID has been “stolen” from z/OS using the HMC/SE CHPID Config OFF/ON facility instead of the z/OS CF CHP(cc).75.IBM. Appendix F. The information displayed is a direct reflection of a path mask called the Path Operational Mask (POM) in the UCW. The PATH ONLINE field indicates whether the path is logically online to z/OS and will change as a result of the VARY PATH command.FICON Director @ 66. Port 88. PATH ONLINE = Y.75. Useful z/OS commands 297 .PATH OPERATIONAL Y MANAGED N CU NUMBER D000 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107. Information for this field is obtained from the UCBLPM (UCB Logical Path Mask). CHP PHYSICALLY ONLINE = Y.0300 SCP TOKEN NED = 002107. ZHPF For both D M=DEV examples. The FICON Director exit port: 6688 . This response can be due to: – – – – Disabling a channel interface Host Bay Adapter (HBA) at the control unit Disabling a port on a FICON Director Powering off a control unit or cluster Powering off a DASD controller Note: The PATH OPERATIONAL status can be easily misinterpreted. The FICON Director entry port: 6503 .IBM.0000000BALB1.IBM.

900. RNID = xxxxx.75.xxx.xx.xx.xxx.IBM.IBM. See Example F-14 on page 299 for a display of an alias device via D M=DEV command.xxxx (not shown in the example) This shows the Node-Element Descriptor read from the device in response to this command.0003 This shows the Node-Element Descriptor last obtained by the SCP. Example: F-13 Using ‘D M=DEV without specifying a device D M=DEV IEE174I 13. # 4 .xxxxxxxxx.IBM.xxxxxxxxxx.xxx. 1 1 1 1 2 2 .xxxxxxxxxx. This RNID information is kept in the Hardware System Area (HSA). HIPERPAV ALIASES CONFIGURED = 143 This field shows the number of (ALIAS) HIPERPAV ALIAS devices configured in this LCU.xxx. . SCP DEVICE NED = 002107.xxx. SCP CU ND = 002107.xx.xx. . FUNCTIONS ENABLED = MIDAW.0300 This shows the Node Descriptor (ND) LAST obtained by the System Control Program (SCP). ACTUAL TOKEN NED = xxxxx.0000000BALB1.CU NUMBER = D000 This field presents the Control Unit Number associated to the physically attached defined CU.75. DESTINATION CU LOGICAL ADDRESS= 00 The DESTINATION CU LOGICAL ADDRESS field shows the logical link address of a control unit that is used to access the I/O device associated with the specified channel paths. ACTUAL CU ND = xxxxxx. # 4 F . The information for each defined device will be provided by the display. . # # # # # 4 4 4 4 4 E .0000 This shows the Node-Element Descriptor last obtained by the SCP.75.0000000BALB1.xxx.xxxx (not shown in the example) This shows the Node-Element Descriptor read from the device in response to this command.900. SCP TOKEN NED = 002107. ZHPF This field shows the additional facilities that this device is capable of using.46.xxx. Using a D M=DEV command without specifying a device Example F-13 shows the output of a display device (D M=DEV) command without specifying a device number. .xxx. including device-specific information about its current state.0000000BALB1.46 DISPLAY M 238 DEVICE STATUS: NUMBER OF ONLINE 0 1 2 3 4 5 6 7 8 0006 . ACTUAL DEVICE NED = xxxxx.xxxxxxxxxx.xxxx (not shown in the example) This shows the Node Descriptor of the attached subsystem read from the device in response to this command. .900. 00B0 # # # # # # # # # 0180 4 4 4 4 4 4 4 4 4 298 FICON Planning and Implementation Guide CHANNEL PATHS 9 A B C D .xxxx (not shown in the example) This is the Control Unit Remote Node IDentifier obtained by the Channel Subsystem during Channel .CU initialization using Store Subsystem Channel Information (CHSC).

.means the path is operational (POM). 4@ indicates an out-of-line path condition. 4 4 4 . where: Online . that is. Operational .means the path is online. .. provides symbols identifying the state of existing devices. Significant symbolic meanings are explained here: DN The DN symbol means that the device has been defined to the z/OS IOGEN (OSCONFIG) but has not been defined in the IOCP. and operational indicators are not equal with at least one of these paths.. . For example. . Physically online . Useful z/OS commands 299 .means the CHPID is physically online.22. 0404 DN DN DN DN 1 1 1 1 DN 04B0 1 1 1 1 . 1 1 1 . 1 1 1 1 1 1 1 EXPLANATIONS ************************ OPERATIONAL INDICATORS ARE NOT EQUAL ..This normally occurs when the wrong IOCDS or OSCONFIG is loaded. 1 0504 DN DN DN DN 1 1 1 1 DN . Example: F-15 Displaying a non-existing device via ‘D M=DEV(dddd) D M=DEV(4000) Appendix F. online. . Example: F-14 Displaying an ALIAS device using D M=DEV(dddd) D M=DEV(D080) IEE174I 18... DOES NOT EXIST SUBCHANNEL NOT AVAILABLE SUBCHANNEL IN PERMANENT ERROR DEVICE IS AN UNBOUND ALIAS HYPERPAV ALIAS UNUSABLE ‘D M=DEV’. 1 Using a D M=DEV(dddd) command for an ALIAS device number Example F-14 shows an output of a D M=DEV(dddd) issued against an ALIAS device address. The numeric digit (1) indicates the number of paths defined for the device. PHYSICALLY ONLINE. 4 4 4 . . Its possible to have a combination of a digit and a symbol. ... which is shown in Example F-13 on page 298. The dot (. AND + ONLINE # DEVICE OFFLINE BX DEVICE IS BOXED SN DN DEVICE NOT AVAILABLE PE AL DEVICE IS AN ALIAS UL HA DEVICE IS A HYPERPAV ALIAS HU 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 .. There is no UCB available for this device in the OSCONFIG being used. . . The UCB for a DN device has the Not Connected bit left on during the device mapping time during IPL.19 DISPLAY M 660 DEVICE D080 STATUS=HYPERPAV ALIAS Using a D M=DEV(dddd) command to a non-existing device Example F-15 on page 299 shows the output of D M=DEV(dddd) issued against a non-existing device. . there is no subchannel for this device known to the channel subsystem. 0FB0 # # # # 1 1 1 1 1 ************************ SYMBOL @ ONLINE.0181 4 4 4 4 4 4 4 4 4 0182 4 4 4 4 4 4 4 4 4 0183 4 4 4 4 4 4 4 4 4 . DN DN DN . physically online. DN DN DN .) indicates that the device has not been defined to the z/OS in the config.

PPRC. in cylinders Device extended function status Unit Control Block (UCB) device type information Help text. With the DEVSERV command you can display: Device Number Device type Logical mode of the device Number of data sets allocated on the volume Volume serial label Channel path ID Status of the path Status of an SMS-managed device – Volume Status – Storage Group name – Storage Group status Control unit type and model Control unit serial number Device capacity. The response is a display of basic status information about a device. and optionally can include a broad range of additional information.dddd.23.n The DEVSERV command requests a display of the current status of DASD and TAPE devices. if the device belongs to the DASD storage subsystem: – Real device type (if what is shown is an emulated device type) – Control unit type and model (or emulated control unit type and model. a group of devices. if the real and emulated control units are not the same) – Subsystem ID for this storage subsystem – Cache fast write state – Track caching state – DASD fast write state – State of pinned data – State of dual copy. or SPARing (if there is any) – Address of the other device in a dual copy pair – Channel subsystem device address – Subsystem internal logical device address 300 FICON Planning and Implementation Guide . or storage control units.IEE174I 18. when you request it The following information can be displayed.DS P.52 DISPLAY M 666 DEVICE 4000 STATUS=NOT IN SYSTEM DEVSERV command .

To learn about other uses of the DEVSERV command. The path status for all defined channel paths is included.53.DPCT . or storage control units.RDC. Example: F-16 Devserv QDASD with options RDC and DCE DS QD. or to DFSMS Storage Administration Reference.TYPE=3490<. and cartridge volume serial number.dddd.1C00 . SA22-7627.DCE IEE459I 09.xxxx.RCD.Display MVS control blocks (1 device).16.SNSS .– An indication if the device extended function status information is inconsistent between z/OS control blocks and the storage subsystem – An indication if the defined (UCB) device type is inconsistent with the real device type – Optionally. The response is a display of basic status information about a device.Test device 1C00 only through all paths. DS QD. DS QPAVS.4800.n . You can use the QTAPE (QT) parameter to display tape information.nnn allows you to display information for the device type.1. 3390. There are different ways of using the DEVSERV command: DS P.1. Appendix F. and optionally can include a broad range of additional information. DS QD.Test only ONLINE devices in 2000-200F.RDC.9000.MED. if the device belongs to a tape library: – Device type equivalent to DTYPE from the DS P command – Device status indicating online/offline and ready/not ready – Device type and model – Device serial number – Library identification number – An indication if the defined (UCB) device type is inconsistent with the real device type Example F-16 shows the output of a DEVSERV command. and 9345) for all of the devices within the scope of the request The following information.DCE. refer to z/OS MVS System Commands. Useful z/OS commands 301 . media type.Display status of Parallel Access Volume devices.3200.Display TAPE with 3490 information. the total number of cylinders for each unique track format (3380.Get hardware information and update control blocks. DS P.2000. SC26-7402. Using the MED (medium) option DS QT.DEFINED> .SSSCB. a group of devices.UCB.04 DEVSERV QDASD 445 UNIT VOLSER SCUTYPE DEVTYPE CYL SSID SCU-SERIAL DEV-SERIAL EF-CHK 9000 IN9000 2107921 2107000 32760 1000 0113-00511 0113-00511 **OK** READ DEVICE CHARACTERISTIC 2107E833900A5E80 FFF720247FF8000F E000E5A205940222 1309067400000000 0000000000000000 24241F02DFEE0001 0677080F007F4A00 003C000000000000 DCE AT V020A2A20 3878807100C445C0 0000000001F0FED8 D8007FF87FF72424 1FF7080000410000 00FC24DC9400F0FE 001F3C1E00078000 0000000000000000 **** 1 DEVICE(S) MET THE SELECTION CRITERIA **** 0 DEVICE(S) FAILED EXTENDED FUNCTION CHECKIN DS QT.ON . as well.

TOTD94. N SIMPLEX 03 03 10017 2107 ************************ SYMBOL DEFINITIONS ************************ O = ONLINE + = PATH AVAILABLE .O. YY.50=.nnn DS P.33.Z18RB1. 2107 is the true device type where the DEVSERV command has been executed. 302 FICON Planning and Implementation Guide .Mount Pending O .000.50=. M=O This field represents the Device UCB status: A .54=.40 DEVSERV PATHS 919 UNIT DTYPE M CNT VOLSER CHPID=PATH STATUS RTYPE SSID CFW TC DFW PIN DC-STATE CCA DDC CYL CU-TYPE D000.58=+ 5C=+ 7E=+ 82=+ 2107 89E0 Y YY.50=.= LOGICALLY OFF.dddd. VOLSER This field indicates the VOLSER of the device pointed to by the DEVSERV command.O. DTYPE = 33909 This is the device type of the device reported by DEVSERV command.33909 .54=.Not Allocated CNT This field indicates the number of data sets allocated on this volume. Most display commands do not actually perform a real I/O operation and report back the state of specific fields in the UCB and UCW (subchannel).Online P .Offline M .33909 . N SIMPLEX 01 01 10017 2107 D002.Z1BRZ1. In this case. Example: F-17 Output of one type of DEVSERV command: DS P.Note: The DEVSERV command is one of the very few commands that actually performs an I/O operation against the device or devices specified. YY.D000.Z18DL2.nnn command.000.4 IEE459I 18.33909 .002.54=. device D000 reported back as a 3390 model 9. Example F-17 shows the DS P. YY. N SIMPLEX 02 02 10017 2107 D003.58=+ 5C=+ 7E=+ 82=+ 2107 89E0 Y YY. PHYSICALLY OFF The fields reported by the DASD DEVSERV command are listed here. RTYPE = 2107 This is the real (true) device type of the device reported by the DEVSERV command.58=+ 5C=+ 7E=+ 82=+ 2107 89E0 Y YY. YY. SSID This field indicates the subsystem ID of the device CU.58=+ 5C=+ 7E=+ 82=+ 2107 89E0 Y YY.54=.Allocated F . N SIMPLEX 00 00 10017 2107 D001.O.Offline Pending N .A.33909 .50=.000.dddd.

Pending Active F .Pinned Data in Cache or NVS .Destage in Progress S .NVS Pending Unavailable .d) C=Device Cache.CU is running in Single Cluster mode (half of the cache in one cluster is not in use) T .Destage failed I .CFW is Suspended with Pinned Data TC (c. current and original Secondary) .DFW .Pending Inactive M .Destage in Progress or has failed S .DFW Deactivate Pending with Destage failed U .NVS Terminated due to error PIN (p) This single character indicates whether pinned data exists: N .DDC Swapped) SPARE (RAMAC: RDC byte 57=33/34 and SNSS byte 36 bit 4.CFW is Inactive S .CFW/DFW allowed S .Terminated due to a Subsystem Cache Error DFW (e.CFW is Active N .Retriable Pinned Data in Cache/NVS .DDC same) PRI-SSEC (Suspended.CFW Cache Fast Write.Battery Defective P .No Pinned Data Y .DDC same) SEC-SDPL (Suspended.Deactivate Pending .Destage in Progress I .DDC Swapped) SEC-SPRI (Suspended.NVS .5=01/10) SPAR-PNDG (Spare Device being copied or has been copied and Copy-Back is pending) SPAR-BRKN (Broken Device copied to Spare device) PPRIMARY (PPRC Primary) PPRI-PNDG (PPRC Primary Pending) PPRI-FAIL (PPRC Primary Fail) PPRI-SUSP (PPRC Primary Suspended) PSECONDRY (PPRC Secondary) PSEC-PNDG (PPRC Secondary Pending) PSEC-FAIL (PPRC Secondary Fail) PSEC-SUSP (PPRC Secondary Suspend) Appendix F.NVS Pending Unavailable/Disable . was Secondary and now Primary .Active/Available N . which indicates the status of the CFW: Y . current and original Primary) .f) E=DASD Fast Write. F=NVS Status These two characters indicate the DASD Fast Write and the NVS status: Y .Pending Inactive . D=Subsystem Cache These two characters indicate the Device Cache and Subsystem Cache status: Y .CFW/DFW temporarily suspended DC-STATE This field reports the Dual Copy current status: SIMPLEX (Simplex: Not Duplex-Pair) PRIMARY (Primary Device of Active Duplex-Pair) SECONDARY (Secondary Device of Active Duplex-Pair) PRI-PNDG (Establish/Copy in Progress as Primary) SEC-PNDG (Establish/Copy in Progress as Secondary) PRI-SDPL (Suspended. was Primary and now Secondary .Inactive/Unavailable F .Active N .Inactive A .Disabled for maintenance (this will override other statuses) P . Useful z/OS commands 303 .DFW is temporarily Suspended with Pinned Data U .

when applicable. CYL This field reports the number of cylinders of the subject device. 304 FICON Planning and Implementation Guide .This indicates the Device-to-Director Connection address. 7) CCA-xx xx .MIRR-OPER|PEND|FAIL (mirroring status per SNSS byte 26 bit 6. DDC-yy yy .This indicates the Channel Connection address: – The xx value should match UNITADD-XX in the IOCP. CU-TYPE This field reports the control unit type of the subject device.

a single FICON CHPID can be used simultaneously to talk to another FICON CHPID and any other native FICON controller. The FICON channel at each end of the FICON CTC connection. You can choose one of the following FICON channel CTC configurations: One FICON channel in a point-to-point. 305 . Using a pair of FICON native channels allows the installation to maintain the same CTC device definition methodology for FICON as was previously used for ESCON. © Copyright IBM Corp. performance levels can be kept at a premium if the channel utilization is kept at recommended levels (under 50%). All rights reserved. 2009. Although a single FICON native channel per server can provide CTC connections across multiple LPARs. at the same time. sending and receiving data concurrently. With the System z. such as disk and tape connected via a FICON Director. such as disk and tape.G Appendix G. In a mixed CTC/controller configuration. The FICON channel CTC communication does not require a pair of channels because it can communicate with any FICON channel that has a corresponding FCTC control unit defined. Because the FICON native channel supports a larger number of devices. can also communicate with other FICON native control units. for large FICON configurations we recommend using at least one pair of FICON native channels. But the FICON channels can support the definition of FCTC control units as well as other I/O control units. 2006. or cascaded FICON Directors configuration Two FICON channels in a point-to-point. installations with a high number of logical partitions in an FCTC complex are easier to design and define. 2005. switched point-to-point. switched point-to-point. supporting the FCTC control units. or cascaded FICON Directors configuration The channel operates in full duplex mode. 2006. Adding FICON CTC connections This appendix contains recommendations for adding CTC connections to a FICON environment.

However. independent of the image and processor where that operating system runs. a further performance enhancement can be made. and evenly spread the receive CU/device definitions across the two FC channels. Defining a FICON CTC configuration requires an understanding of CTC operations and the different methods available for defining the connections. operations and systems programmers can easily identify the use (send or receive) and target system for any given CTC device. define all send FCTC control units and devices on one FICON (FC) channel. Fourth digit This digit is used to indicate whether the CTC connection is the primary or alternate (backup for availability purposes). A value of 0 to 7 is used for the primary CTC connection devices. This CTC image-ID is used as a target identifier (for the image) that you use when defining how all other images access this image. Apart from ease of migration. refer to: FICON CTC Implementation. evenly spread the send CU/device definitions across the two FC channels. and a value of 8 to F is used for the alternate CTC connection devices. assigned on paper) for the LPAR image. and can be any unique value within the CTC complex. Second and third digits These digits represent an assigned CTC image-ID (that is. It also simplifies the CTC definition process in that the same operating system CTC device definitions can be used by all z/OS operating systems. the CTC definition process becomes increasingly complex. For more details. this requires a large amount of customization and may not be practical with very large configurations. To fully exploit the full duplex data flow capabilities of the FICON channels. It is also necessary to take into consideration the subsystem requirements for using CTC connections. REDP-0158 306 FICON Planning and Implementation Guide . Using two FICON (FC) channels per server. and define all receive FCTC control units and devices on a second FC channel.FCTC device numbering scheme There are a number of schemes that can be used for developing the device numbers for the FICON CTC devices. It also makes it difficult to relocate operating systems between logical partitions on the same server or a different server. – An odd hexadecimal number is used for the receive CTC control unit and device. Using the FICON sender/receiver device numbering scheme is the recommended approach. The method makes use of the 4-digit z/Architecture device number where: First digit – An even hexadecimal number is used for the send CTC control unit and device. We do not recommend using the matching device number scheme because it does not enable the operator or systems programmer to identify the use or target system for a given CTC device number. As the number of images in a complex grows. Using two FICON channels per server.

SG24-7543 IBM System Storage/Brocade Multiprotocol Routing: An Introduction and Implementation. SG24-6857 Getting Started with the INRANGE FC/9000 FICON Director. REDP-4187 How Does the MIDAW Facility Improve the Performance of FICON Channels Using DB2 and Other Workloads?. SG24-7544 Implementing an IBM/Cisco SAN. 2005. IBM Redbooks For information about ordering these publications. SG24-5444 Getting Started with the IBM 2109 M12 FICON Director. GA22-7589 © Copyright IBM Corp. GA22-7588 z/OS MVS Diagnosis: Tools and Service Aids. Note that some of the documents referenced here may be available in softcopy only. SG24-6089 Implementing an IBM/Brocade SAN with 8 Gbps Directors and Switches. SG24-6397 FICON Implementation Guide. IBM System z Connectivity Handbook. SG24-7545 IBM System z10 Enterprise Class Configuration Setup. All rights reserved. REDP-4387 Cisco FICON Basic Implementation. 2009. SG24-7571 FICON CTC Implementation. REDP-4201 Multiple Subchannel Sets: An Implementation View. SG24-7308 IBM System z10 Enterprise Class Technical Introduction. SG24-7515 IBM System z10 Enterprise Class technical Guide. FCIP. 307 . and FICON Environment. see “How to get Redbooks” on page 309. REDP-4392 Other publications These publications are also relevant as further information sources: z/OS MVS Diagnosis: Reference. SG24-6786 Getting Started with the McDATA Interpid FICON Director. SG24-6858 IBM Tivoli System Automation for z/OS Enterprise Automation. 2006. REDP-0158 Disk Storage Access with DB2 for z/OS. SG24-6116 Implementing the Cisco MDS9000 in an Intermix FCP. SG24-6497 IBM System Storage DS8000 Architecture and Implementation. SG24-7516 IBM/Cisco Multiprotocol Routing: An Introduction and Implementation.Related publications The publications listed in this section are considered particularly suitable for a more detailed discussion of the topics covered in this book.

53-1001185 High Performance FICON for System z Technical Summary for Customer Planning. GA52-1333 IBM System Storage DS8000 Introduction and Planning Guide. GC35-0515 FICON Express2 Channel Performance Version 1. ZSW03058USEN Performance Considerations for a Cascaded FICON Director Environment Version 0. and Open System Adapters). InfiniBand. SC33-7990 z/OS RMF Report Analysis. TSD00069USEN Cisco MDS 9509 for IBM System Storage. 9509. SC28-6879 HCD User’s Guide. Service. SC33-7992 z/OS RMF Programmer’s Guide. SC28-6830 HMC Operations Guide. TSD01754USEN Cisco MDS 9506. ZSW03059USEN IBM System Storage SAN768B. and Open System Adapters). SY27-2597 IBM Fiber Optic Cleaning Procedure.2x. GA23-0367 IBM System Storage SAN768B Installation. SA22-1084 z/OS MVS System Commands. GM13-0702 System z10 Enterprise Class System Overview. and User's Guide.System z Planning for Fiber Optic Links (ESCON. SA22-7832 System z Input/Output Configuration Program User’s Guide for ICP IOCP. SC28-6873 System z10 Support Element Operations Guide.pdf IBM System z9 I/O and FICON Express4 Channel Performance.com/servers/eserver/zseries/library/techpapers/pdf/gm130237. and User's Guide. SX33-9033 Brocade Fabric OS Administrator’s Guide. SC33-7994 System z Maintenance Information for Fiber Optic Links (ESCON. Coupling Links. SA22-7627 z/OS MVS System Messages SA22-7637 z/Architecture. SC26-7402 HMC Operations Guide. SY27-2604 z/OS RMF Reference Summary. SC33-7988 z/OS RMF User’s Guide. ZSW03005USEN IBM System z10 I/O and High Performance FICON for System z Channel Performance. 9513 for IBM System Storage Directors Interoperability Matrix 308 FICON Planning and Implementation Guide .ibm. SB10-7037 System z10 Processor resource/Systems Manager Planning Guide. Principles of Operation. TSD00070USEN Cisco MDS 9513 for IBM System Storage. Coupling Links. Service. SB10-7153 DFSMS Storage Administration Reference. FICON. TSD03037USEN IBM System Storage SAN768B Fiber Backbone Interoperability Matrix Cisco MDS 9506 for IBM System Storage. SC33-7991 z/OS RMF Performance Management Guide. Richard Basener and Catherine Cronin http://www-03.0. GA32-0574 IBM System Storage SAN384B Installation. FICON.

draft publications and Additional materials.com/redbooks Help from IBM IBM Support and downloads ibm. Inc.brocade. Web site http://www.Online resources These Web sites are also relevant as further information sources: Fibre Channel standard Web site http://www. Technotes.org Brocade Communications Systems.com/support IBM Global Services ibm. or download Redbooks. Web site http://www.cisco.com How to get Redbooks You can search for.com Cisco Systems. view. Inc.t11. at this Web site: ibm. Redpapers. as well as order hardcopy Redbooks publications.com/services Related publications 309 .

310 FICON Planning and Implementation Guide .

133. 26. 202. 143. 99. 135. 277. 8. 130. 103. 29 CHPIDS 12. 14. 26 CHPID 100. 2005. 99. 144.CHPI D 2. 288 CF CHP 109. 115. 139 CNTLUNIT statement 104. 139. 133. 246 DCFM server 153. 116. 246 IP address 162 © Copyright IBM Corp. 272. 148 channel path functional details 127 shared definition 122 channel path (CP) 11. 26 FCP 26 FCV 9. 149. 135. CUs (CPC) 138. 152–155. 226. 120. 257 CNTUNIT statement 122 command mode 14 CCW operation 20 channel program 21 command response (CMR) 204 CONFIG command 108. 185. 268 Cisco MDS 9506 44 Cisco MDS 9509 44 Cisco MDS 9513 44 CNTLUNIT CUNUMBR 104. 124. 289 C Cache Fast Write (CFW) 302 cascaded FICON Director Additional configuration steps 164 CCW 17 other parameters 17 CCW chaining 204 CCW execution 20 control unit 20 Central Processor Complex (CPC) 107.Index A active CSS 108. 255. 138. 206. 147 2. 123. 119. 120. 218. 2006. 153 Control unit logical link address 298 logical-path establishment 14 login 14 node-identifier acquisition 14 serial number 300 state-change registration 14 type 300 control unit 271 N_Port address 39 node descriptor 24 control unit (CU) 1. 147 D DASD Fast Write (DFW) 302 Data Center Fabric Manager (DCFM) 47. 140 Cyclical Redundancy Checking (CRC) 41 B buffer credit 23. 126. 14. 124. 125. 311 . 255 CHPID number 113. 295 Control Unit Port (CUP) 39. 273 CSS.7E 129. 255. 279–281. 257 CHPID statement 38. 143 connecting FICON Director Domain ID 42 control processor (CP) 45. 120. 290 Active Zone Config 188 Advanced Performance Tuning Policy 176 American National Standard Institute (ANSI) 4 attached FICON Director error condition 44 CHPID Mapping Tool (CMT) 103. 211 build IOCDS 107. 297 changing passwords on Director 162 Channel feature 26 Channel Information PD panel 272 window 112. 2009. 275. 257 CHPID type FC 10. 152 control units as specified 138 D000 120. 139 CPC icon 268 CRC errors 91 CSS 2 103. 138.7F 111 CU 4044 105. 270. 120. 34. 273 current status 126 direct connection 105 CHPID detail 111. 209. 284 Channel Subsystem channel path 257 logical control units 291 logical partition 289 not-operational reason 237 Channel Subsystem (CSS) 288 Channel-to-Channel (CTC) 11. 220. 39. 138. 125. 220. 235. 266. 120. 129. All rights reserved. 131. 255. 266 Channel Path Identifier 12 channel paths. 142 Business Class (BC) 28. 116. 119. 295 CHPID 7E 103. 219 Channel Problem Determination 272. 115. 229. 103. 138. 152.

244. 305 arbitrary number 257 Basic functions 37 control unit 35 cooling capability 37 CP cards 154 CUP port 121 default buffer credit assignment 248 detailed configuration 210 Domain ID range 38 fabric binding database 42 FC links 35 Fiber optic cables 144 IBM qualification testing 44 in-band management 39 ISL port 190 Other settings 125 physical port 11 port address 185 switch module 37 Switch number 282 FICON environment 4. 199. 306 option Search 277 Device Status (DS) 109. 43. 13. 115. 287. 292 cam display information 301 DEVSERV command 292 other uses 301 Devserv command 300 Director port (DP) 208 Domain ID 37. 133. 13. 220–221. 257 DS P 225. 271 CHPID type 257 different flavors 203 fiber links 108 full duplex data flow capabilities 306 IOS051I message 232 PEND time ends 204 physical connectivity 37 physical transmission path 36 topologies 34 FICON channel (FC) 210 FICON Channel-to-Channel (FCTC) 121. 125. 135 Feature code 26. 116. 137 device 100. 57. 202 FICON 1. 14. 16. Editing. 179. 101–102. 140 FICON Director 1. 233 F F_Port 6 fabric 6 Fabric Manager (FM) 54 fabric port (F_PORT) 6. 305 FICON advantage 9 FICON channel 5. 133. and Printing (EREP) 222 ESCON Director 39 ESCON solution 9 Establish Logical Path (ELP) 15 Extended Link Service (ELS) 19 Extended Subchannel Logout Data (ESLD) 223. 99. 289 Enterprise Fabric Connectivity Manager (EFCM) 155 Environmental Record. 33. 305 Feature Check 100. 128. 257. 134 system 100. 217. 54. 244. 255. 275. 152 Dynamic Path Selection (DPS) 49. 116. 247. 145. 34. 265. 140. 224. 222 Protocol traffic 26 Routing 48 Security Protocol 54 Single Byte Command Sets-3 13 Single Byte Command Sets-4 13 Single-Byte-3 8 Single-Byte-4 8 standard 4 Switch Fabric and Switch Control Requirements 8 Fibre Channel (FC) 3. 33–34. 136 FICON Express4 10KM LX 26 10KM LX feature 27 4 KM LX feature 28 4KM LX 26 CHPIDs 27 312 FICON Planning and Implementation Guide . 133. 268. 220. 118. 145. 119. 19. 99. 183 Enterprise Class (EC) 244. SB-4 Web site 5 FCTC control unit 11. 203. 292. 46. 3. 185. 300 DS8000 storage control unit 108. 152. 15. 257. 218. 39 bit error 223 data traffic 55 fabric 39 link 18 logical switch 197 Physical 4 Physical and Signaling Standard 8 physical framing 5 physical interface 4 port 222 Protocol 10. 249 fiber optic cable 6. 10. 257. 305 FICON Express 30 FICON Express2 3. 118. 134 Dynamic Load Sharing (DLS) 49. 233 Dynamic Pathing Switch (DPS) 231 E E_Port 6. 136. 34. 104. 116. 197. 16. 101. 251. 293 device type 110. SB-3. 38. 34. 115. 297. 35 FC-SB-2. 280. 150–152. 201. 151. 97. 179 FC link 6. 225. 146. 127. 135. 116. 175.ODBC password 162 welcome window 162 defined channel path path status 301 Device Number 109. 143 controller 102. 272. 127. 134–135 CHPID 82 105 Fibre Channel adapter 17 Arbitrated Loop 6 architecture 1. 105.

235–236. 142.5 30 mode conditioning patch (MCP) 27–28 MULTIPLE IMAGE FACILITY (MIF) 288 Multiple Image Facility (MIF) 12 N N_Port 4. 152 Input Output Configuration Program (IOCP) 8. 126 Insistent Domain (ID) 152 Inter-Chassis Link (ICL) 48 Inter-Switch Link (ISL) 35. 117. 229. 181. 138. 154 Input/Output architecture 11 Input/Output Supervisor (IOS) 16. 289 HCD User 103. 34. 124. 289 Hardware System Area (HSA) 14. 256 I/O definition file (IODF) 103. 37. 256 Host Bay Adapter (HBA) 221. 122. 137 file 106–107. 245 K km transceivers 28 H Hardware Configuration Definition (HCD) 8. 244 FICON FC-SB3 and SB-4 Web site 19 FICON feature 8. 257. 138. 220. 42. 260 ISL Trunking 54 ISLs 35. 291 logical partition channel subsystem 292 high number 305 logical partition (LP) 12. 26 SX 26 FICON Express4 10KM LX 27 FICON Express4 4KM LX 28 FICON Express4 SX 28 FICON Express4-2C 4 KM LX 28 FICON Express4-2C SX 28 FICON Express8 3. 119. 16. 206. 247. 298 unit addresses 258 I/O information 206 I/O operation 12. 256 M MIF ID 257. 100. 14. 203. 38. 133 Input/output (I/O) 3. 119. 290 IOS 220 IOS command 108. 101. 135 Link-Incident-Record Registration (LIRR) 14 logical control unit (LCU) 100. 207. 289 Program 255. 51. 272. 135. 288 CU 5034 105 LPARs 104. 247 FICON Purge Path Extended 222 FL_Port 6 G G_Port 6 Gbps 16. 231. 265. 221. 289 I/O definition 103. 175 Invalid Transmission Words 91 IOCP User 256 IODEVICE Address 104. 38. 206. 141. 122. 258 PARTITION keyword 105 IODF data (ID) 107. 123. 119. 289 Missing Interrupt Handler (MIH) 21. 35. 255. 219. 117. 209. 281. 247 management isolation 50 port numbering 173 logical switch types 50 Long Wave (LW) 46 Longitudinal Redundancy Checking (LRC) 41 longwave laser 101. 38.feature 9. 244 graphical user interface (GUI) 54. 267. 140. 282. 115. 275. 233 MM 62. 143 IP address 55. 154. 119. 292 exchange pair 19 FICON architecture 16 MIH value 22 I/O performance measurement points 203 I/O priority queuing (IOQ) 213 I/O processor (IOP) 209 I/O request 16. 121. 119. 134 I I/O Configuration Data 255 Data Set 99.Partition Brocade SAN768B 50 logical switch 50. 255. 105–106. 128. 146. 289 Hardware Configuration Manager (HCM) 255. 134–135 Loss of Sync 91 lossless DLS 175. 133. 289. 119. 246. 248 LPAR A23 105. 125. 276 Logical switch scalability. 135. 289 FC-FS FC-2 frame 18 I/O specification 103. 273. 117. 139 IODEVICE statement 103. 152. 170. 138 I/O device 11. 152. 138 IBM RMF Web site 203 IBM SAN384B 44 IBM SAN768B 44 IBM System Storage Web site 93 in-order delivery (IOD) 49. 152. 103. 138. 152. 266 Hardware Management Console 110. 116. 13. 141. 14 Index 313 . 17. 115. 255. 134. 297 HYPERPAV ALIAS (HA) 299 L L_Port 6 LC Duplex 27 LC Duplex connector 27. 136. 305 Logical Path 15. 135. 99. 116. 13.

144 Transport Control Word (TCW) 16 Transport Indirect Data Address Word (TIDAW) 22 transport mode 14 I/O operation 14 Trivial file transfer protocol (TFTP) 43 trunk or ISL (TI) 152 Q QoS SID/DID pairs 49 Quality of Service (QoS) 48. 161. 222 Store System Information (STSI) 288 subchannel 276. 116. 222. 212 Path Available Mask (PAM) 277 PATH keyword 103. 261 port type 5. 192 switched fabric 11 System Activity Display (SAD) 201 System Control Program (SCP) 11. 279 Subchannel Logout Handler (SLH) 223 subchannel number 275 subchannel set 12. 297 PCHID 5A2 116. 146. 120. 135. 285 T Tag 274 TE port 54 terminology 4 TI zone 190 Traffic Isolation (TI) zone 48 Traffic Isolation Routing feature 48 transceiver 126. 217. 138 PPRC Secondary Suspend (PSEC-SUSP) 303 problem determination (PD) 108. 138. 39. 197. 273 Essential information 129 owning image 129 PCHID 5A3 100 Essential information 111 owning image 111 PCHID detail 110. 280 port fencing 91 port number 38. 174. 100. 13. 163. 289 Systems Automation (SA) 39 O operating system 17. 214 Port-Based Routing (PBR) 152 Power-on Reset (POR) 103. 149. 127. 265 Processor Card (PC) 160 Purge Path Extended (PPE) 222. 101. 100. 144. 99. 298 System Information (SI) 288 System Management Facility (SMF) 203 System z 3. 210 average number 210 NL_Port 6 node 6 Node Descriptor (ND) 16. 134 Single Object Operation (SOO) 110. 266 Small Form Factor Pluggable (SFP) 37 State-Change Registration (SCR) 14 Storage Area Network (SAN) 43 storage device 34. 275 Unit Control Block (UCB) 17. 244. 100. 255. 13 performance monitoring 203 point-to-point configuration 11. 100. 257 Path Not Operational Mask (PNOM) 231 Path Operational Mask (POM) 231. 179. 305 High Performance FICON 26 System z environment 8 System z server 134 System z10 8. 139. 134. 255 FICON environment 99 port 6 port address 8. 247. 258. 138. 128. 37. 128. 277. 257 point-to-point topology 35. 134. 244. 117. 100. 306 FICON Express support 137 FICON Express8 support 102 Operation Request Block additional parameters 17 Operation Request Block (ORB) 17 P Parallel Access Volume (PAV) 8. 108. 135. 152 U UCB Logical Path Mask (UCBLPM) 297 unit address (UA) 9. 18. 255.Native FICON 26. 117. 295. 117. 298 Resource Measurement Facility (RMF) 202–204 314 FICON Planning and Implementation Guide . 177. 126. 288 Node-Element Descriptor (NED) 296 RMF 204 S SCP CU Neodymium 110. 300 Unit Control Word (UCW) 291 Unrepeated distance 9. 146 PCHID number 270 PCHIDs 268 performance FICON 9. 38. 265 switch address 11. 237. 116. 125. 120. 277. 152. 135. 257 Switch Connection Control (SCC) 153. 103. 120. 115. 13. 33. 119. 289 Support Element (SE) 240. 143. 146. 274. 116. 16. 27 Upper Level Protocol (ULP) 5 R Redbooks Web site 309 Contact us xiii Registered State Change Notification (RSCN) 177 Remote Node Identification (RNID) 274. 185. 296 Secure File Transfer Protocol (SFTP) 54 selected central processor complex data sets 256 Setting the display to hex 163 Short Wave (SW) 46 Simple Network Management Protocol (SNMP) 37 single mode (SM) 47. 116. 266.

134 configuration tasks 110 corresponding LPAR 140 designated ports 128 fiber optic cable link 113. 113. 170 Virtual Fabrics feature 50 Virtual ISL 49 Virtual Storage Area Network (VSAN) 54 W Wave Division Multiplexing (WDM) 196 working CHPID normal indication 271 World Wide Name 7 World Wide Node Name 16. 30. 28. 117. 131. 131. 149 FICON Express8 feature 102 PCHID 5A2 128 PCHID 5A3 110 standard feature 100 zHPF feature 101. 193 World Wide Node (WWN) 113. 149. 135. 291 z10 EC 12. 291 FICON connectivity 247 z10 server 100. 149.User Interface (UI) 266 V Virtual Fabric 51. 116. 116. 222 World Wide Port (WWP) 113. 218 Zone Config 187 Index 315 . 149. 113. 248 zHPF on System (Z/OS) 100. 149 World Wide Port_Name (WWPN) 7 Z z10 BC 12. 134. 131. 247. 169 World Wide Node Name (WWNN) 42 World Wide Node_Name (WWNN) 7 World Wide Port Name 16. 131.

316 FICON Planning and Implementation Guide .

5” spine) 0.FICON Planning and Implementation Guide FICON Planning and Implementation Guide FICON Planning and Implementation Guide (0.475”<->0.873” 250 <-> 459 pages FICON Planning and Implementation Guide .

FICON Planning and Implementation Guide FICON Planning and Implementation Guide .

.

as well as a broad understanding of the Fibre Channel and FICON architectures. and FICON Directors. The target audience for this document includes IT Architects. FICON Channel-to-Channel (FCTC). It discusses the FICON and Fibre Channel architectures. The book focuses on the hardware installation and the software definitions that are needed to provide connectivity to FICON environments. Specific recommendations are provided to help you implement IT solutions more effectively in your environment. implementation. It also discusses utilities and commands that are useful for monitoring and managing the FICON environment.Back cover ® FICON Planning and Implementation Guide ® Topologies. HCD or IOCP. The reader is expected to have a basic understanding of IBM System z10 and IBM System z9 hardware. concepts. and migration guidance Realistic examples and scenarios This IBM Redbooks publication covers the planning. and supported topologies. terminology. Experts from IBM. and system programmers who plan for and configure FICON environments. implementation. and terminology Planning. and management of IBM System z FICON environments.com/redbooks SG24-6497-02 ISBN 0738433284 . Customers and Partners from around the world create timely technical information based on realistic scenarios. SAN administrators. For more information: ibm.You will find configuration examples required to support FICON control units. INTERNATIONAL TECHNICAL SUPPORT ORGANIZATION BUILDING TECHNICAL INFORMATION BASED ON PRACTICAL EXPERIENCE IBM Redbooks are developed by the IBM International Technical Support Organization. data center planners.

Sign up to vote on this title
UsefulNot useful