You are on page 1of 10

VERITAS Cluster Server

Application Note for Sun StorEdge A3x00 Disk Array


Solaris

August 2001 30-000338-011

Disclaimer The information contained in this publication is subject to change without notice. VERITAS Software Corporation makes no warranty of any kind with regard to this manual, including, but not limited to, the implied warranties of merchantability and tness for a particular purpose. VERITAS Software Corporation shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this manual. Copyright Copyright 2000-2001 VERITAS Software Corporation. All rights reserved. VERITAS is a registered trademark of VERITAS Software Corporation in the US and other countries. The VERITAS logo and VERITAS Cluster Server are trademarks of VERITAS Software Corporation. All other trademarks or registered trademarks are the property of their respective owners. Printed in the USA, August 2001. VERITAS Software Corporation 350 Ellis St. Mountain View, CA 94043 Phone 6503358000 Fax 6503358050 www.veritas.com

VERITAS Cluster Server Application Note: Sun StorEdge A3x00

Introduction
VCS operates with the Sun StorEdge A3000 and A3500 (A3x00) family of disk arrays and RAID Manager 6.22.VCS supports the StorEdge A3x00 in a dual-bus conguration.This document provides instructions for installing and conguring the StorEdge A3x00 with VCS.

Hardware Requirements
x x x x x x

One or more (20 maximum) StorEdge A3x00 disk arrays. A3000: Two Sun Ultra E450, Ultra E3000, Ultra E4000, Ultra E5000, Ultra E6000, or Ultra E10000 servers. A3500: Two Sun Enterprise 2, Ultra E250, Ultra E450, Ultra E3000/E3500, Ultra E4000/E4500, Ultra E5000/E5500, Ultra E6000/E6500 or Ultra E10000 servers. Four UDWIS SBus cards for each StorEdge A3x00. PCI: Two Dual-Channel Ultra SCSI Host Adapters for each StorEdge A3x00. Four differential SCSI cables for each StorEdge A3x00.

Software Requirements
x x x

Solaris 2.6, 2.7, or 2.8. VCS 2.0 or later. RAID Manager 6.22, revision 01.14 or later.

Required Patches

Required Patches
The following required patches are currently included with the RAID Manager 6.22 media:
x x

Solaris 2.6: 105181-15, 105357-04, 105580-13, 105797-06, 106226-01, 107280-03, 105356-09, 105375-15, 105600-07, 106125-07, 106655-02 Solaris 7: 106541-05, 107292-02, 107469-01, 107473-01, 106552-03, 107458-02, 107472-01, 107474-01

Setting Up the Hardware


1. Congure VCS on all servers as described in the VERITAS Cluster Server Installation Guide. 2. Verify that all hardware is properly installed. (If necessary, refer to Hardware Requirements on page 1.) 3. Change the host SCSI bus address (target ID) on one of your servers to an unused number. Note Do not use the numbers 4 and 5 for the target ID; the two controllers for the StorEdge A3x00 use the target ID numbers 4 and 5.

VCS Application Note: Sun StorEdge A3x00 Font>

Setting Up the Hardware

4. Connect the array to the UDWIS adapters located on each VCS server. Connect one adapter from each host to Controller A, and one to Controller B, as shown in the gure below.

UDWIS

UDWIS

c1 c1

c2

Controller A

Controller B

UDWIS

UDWIS

c1

c2

ServerA

VCS ServerB ServerB

StorEdge A3x00

Dual-Bus Conguration with the StorEdge A3x00

Installing the Software

Installing the Software


1. Verify that all required patches are installed on each host. On each host, type the following command:
# showrev -p

2. Install the RAID Manager 6.22 software packages (listed below) as described in the RAID Manager documentation: SUNWosavcs SUNWosar SUNWoasu SUNWosamn

3. From one of the servers, use RAID Manager (rm6) Maintenance&Tuning->Options-> Firmware Level to upgrade rmware on all A3x00 controllers, to level 03.01.02 or greater. 4. Reboot each system. On each system, type the following command:
# shutdown -y -i6 -g0

The output at the end of the boot process conrms proper installation. Output resembles:
The NVSRAM settings of controller c1t5d0s0(1T63250533) are correct The NVSRAM settings of controller c2t4d4s0(1T63250546) are correct

VCS Application Note: Sun StorEdge A3x00 Font>

Configuring the StorEdge A3x00

5. Verify that the A3x00 is installed and operating. On each server, type the following two commands: a. # lad Output resembles:
c1t5d0s0 1T62549100 LUNS: 0 c2t4d1s0 1T63852894 LUNS:

b. # healthck -a Output resembles:


Health Check Summary Information A3x00:Optimal healthck succeeded!

Conguring the StorEdge A3x00


Congure the Logical Unit Numbers (LUNs) before you use the A3x00. Use the RAID Manager graphical interface (rm6) to monitor and congure the A3x00. Refer to the appropriate RAID documentation for detailed information about the RAID Manager and its functions. When you create LUNs on one system, you must use the boot -r command to reboot the remote systems before the LUNs become visible.

Conguring VCS
LUNs that are created on the A3x00 can be used to create diskgroups and volumes with VERITAS Volume Manager. Congure A3x00 LUNs as Disk resources, or DiskGroup resources if you use Volume Manager.

NFS Failover Setup


For details on reconciling major and minor numbers, see the VCS Installation Guide.

StorEdge A3x00 Maintenance Procedures

StorEdge A3x00 Maintenance Procedures


SCSI termination issues must be considered when one node is connected to the StorEdge A3x00 while the other adapters are disconnected and not terminated. Therefore, the following procedures must be performed before you can disconnect one server from the A3x00 for maintenance while the other server remains active.

Disconnecting the Server


Perform the following steps to disconnect a clustered server from an A3x00: 1. Use the VCS GUI or the hagrp -offline command so that all groups on the server to be disconnected are OFFLINE, and ONLINE on another server in the cluster. 2. Halt the server you want to disconnect for maintenance. 3. Start rm6 on the server that is currently active and prepare to bring A3x00 controllers online and ofine: a. Select Recovery. b. Select the RAID Module to be disconnected. Options->Manual Recovery-> Controller Pairs remains disabled until a RAID Module is selected. c. Select Options->Manual Recovery->Controller Pairs. 4. Use rm6 to bring controller A ofine by select controller A, clicking Place Ofine, and waiting until the utility has responded that the controller is ofine. 5. Disconnect the SCSI cable from Controller A and install a standard SCSI terminator over the disconnected controller adapter. 6. Use rm6 to bring controller A online by selecting controller A, clicking Place Online, and waiting until the utility has responded that the controller is online. 7. Use rm6 to bring controller B ofine by selecting controller B, clicking Place Ofine, and waiting until the utility has responded that the controller is ofine. 8. Disconnect the SCSI cable from Controller B and install a standard SCSI terminator over the disconnected controller adapter. 9. Use rm6 to bring controller B online by selecting controller B, clicking Place Online, and waiting until the utility has responded that the controller is online. You can now perform maintenance on the disconnected server.
6 VCS Application Note: Sun StorEdge A3x00 Font>

StorEdge A3x00 Maintenance Procedures

Reconnecting the server


Perform the following steps to reconnect a clustered server to an A3x00 after maintenance: 1. Halt the server that is to be reconnected. 2. Start rm6 on the server that is currently active and prepare to bring A3x00 controllers online and ofine: a. Select Recovery. b. Select the RAID Module to be disconnected. Options->Manual Recovery->Controller Pairs remains disabled until a RAID Module is selected. c. Select Options->Manual Recovery->Controller Pairs. 3. Use rm6 to bring controller A ofine by selecting controller A, clicking Place Ofine, and waiting until the utility has responded that the controller is ofine. 4. Replace the SCSI terminator with a cable connected to the inactive server. 5. Use rm6 to bring controller A online by selecting controller A, clicking Place Online, and waiting until the utility has responded that the controller is online. 6. Use rm6 to bring controller B ofine by selecting controller B, clicking Place Ofine, and waiting until the utility has responded that the controller is ofine. 7. Disconnect the SCSI cable from Controller B and install a standard SCSI terminator over the controller adapter that has been disconnected. 8. Use rm6 to bring controller B online by selecting controller B, clicking Place Online, and waiting until the utility has responded that the controller is online. 9. After you have reconnected both cables, boot the inactive server. 10. When the server boots, use the VCS GUI or hagrp commands to bring groups back online on the reconnected server.

Troubleshooting

Troubleshooting
w

Only one controller appears after a reboot. Check the controllers activity lights on the front of the A3x00.
x

An amber light indicates a failed controller. To determine or change the state of an A3x00 controller, use rm6 Recovery->Options->Manual Recovery->Controller Pairs. If this is unsuccessful, replace the controller.

Perform the following steps at the command line, or use rm6: 1. Determine if both controllers are active by typing:
# rdacutil -i command

2. If one controller is inactive, type:


# rdacutil -m 2 raid_controller

3. If one controller failed, type:


# rdacutil -U active_controller

To restore the controller, use rm6 Recovery->Recovery Guru. If this is unsuccessful, replace the controller.
w

Excessive RAID console messages and SCSI warnings. The A3x00 disk array monitor daemon (arraymon) passes some event log messages to syslog that also appear on the console. These messages do not indicate problems and resemble the following example:
Jun 21 12:40:27 beta raid: AEN event Host=beta Ctrl=1T63250546 Dev=c1t4d0s0 Jun 21 12:40:27 beta raid: AEN event Host=beta Ctrl=1T63250546 Dev=c1t4d0s0 Jun 21 12:40:27 beta raid: ASC=95 ASCQ=02 FRU=00 LUN=00 LUN Stat=00 Jun 21 12:40:27 beta raid: ASC=95 ASCQ=02 FRU=00 LUN=00 LUN Stat=00 Jun 21 12:40:27 beta raid: Sense=70000600000000980000000095020000000000000000000000000 00000000000000000000000818000000000000000000000000000F0531 54363332353035343620202020202002041D00000000000000000000000 00000000000000000000000

When A3x00 controllers fail, Solaris issues a series of SCSI warning messages. This is normal behavior during failure of SCSI controllers; disk operation by the A3x00 should continue as expected.
8 VCS Application Note: Sun StorEdge A3x00 Font>

You might also like