You are on page 1of 31

Implementing ABARS

Part II of III ABARS Sessions


Summer SHARE
Chicago

000-0193-02 SHARE Session 3035 1


Colleen Gordon
Scott Thomas
Mainstar Software

000-0193-02 SHARE Session 3035 2


Agenda
• Introduction: Application-centric
Disaster Recovery
• ABARS Overview

000-0193-02 SHARE Session 3035 3


ABARS Procedure
• ABARS is executed as a secondary
address space under DFSMShsm and
requires its own procedure:
Usually kept in SYS1.PROCLIB
//DFHSMABR PROC
//DFHSMABR EXEC PGM=ARCWCTL,REGION=5M,TIME=NOLIMIT
//MSYSOUT DD SYSOUT=C
//MSYSIN DD DUMMY
//SYSUDUMP DD SYSOUT=*

000-0193-02 SHARE Session 3035 3


ABARS Procedure
• Add the ABARS procedure name to the
RACF (or OEM) started task table

Most likely requires a RACF admin


unless you have RACF SPECIAL
attribute.

000-0193-02 SHARE Session 3035 3


ABARS High Level Qualifier
• Default is HSMACT
• Is the first qualifier of HSM activity logs
• If you must change it, an IBM patch
command is required
• Create a catalog alias to support the
activity log name

000-0193-02 Share Session 3035 5


ABARS High Level Qualifier
• Define the HLQ as a functional RACF group
allowing the ABARS procedure name, Storage
Management, and possibly Technical Services
with ALTER authority
• Assign SMS management class and storage
group to appropriate volumes with appropriate
migration criteria - no need to delete unless of
version prior to 1.4
000-0193-02 SHARE Session 3035 5
DFSMShsm Parmlib
• Parmlib requires ABARS SETSYS
commands to enable ABARS and set
default vaules
• You must have ALTER authority to
HSM parmlib to update it or provide
values to appropriate support personnel

000-0193-02 SHARE Session 3035 5


DFSMShsm Parmlib
• Identify the name of the ABARS procedure:
SETSYS ABARSPROCNAME(DFHSMABR)
• Identify activity log message level full or reduced
(recommend full)
SETSYS ABARSACTLOGMSGLVL(FULL)
• Identify where activity log is sent, DASD or SYSOUT
(recommend DASD)
SETSYS ABARSACTLOGTYPE(DASD)

000-0193-02 SHARE Session 3035 6


DFSMShsm Parmlib
• Identify whether output tapes will be stacked or
nostack (recommend stack)
SETSYS ABARSTAPES(STACK)
• Identify the name of the ABARS tape unit
SETSYS ABARSUNITNAME(3490)
Esoterics such as TAPEC must be placed in the
SETSYS USERUNITTABLE command
• Indicate the name of the ML2 unit to use during
ARECOVER
SETSYS ARECOVERML2UNIT(3490)
000-0193-02 SHARE Session 3035 6
DFSMShsm Parmlib
• Identify ARECOVER unit name to be used as default
(can be overridden on ARECOVER command)
SETSYS ARECOVERUNITNAME(3480)
• Identify the number of default I/O data movement
buffers (default is 1 recommend can indicate up to 9)
Used by ABARS when HSM is used to read or write
data sets during aggregate backup and recovery
SETSYS ABARSBUFFERS(6)

000-0193-02 SHARE Session 3035 6


DFSMShsm Parmlib
• Indicate if ABARS activity logs should be automatically
deleted during ABARS processing or not
SETSYS ABARSDELETEACTIVITY(Y)
• Set the performance of backing up level 0 DASD data
sets (default is 3, recommend 4)
SETSYS ABARSOPTIMIZE(4)

000-0193-02 SHARE Session 3035 6


DFSMShsm Parmlib
• Indicate if the volcount parameter is to be passed to
DFDSS during aggregate recovery for level 0 DASD
data sets.
Options are Any, NONE or 1 to 59
Default is NONE (don’t pass the VOLCOUNT parm)
ANY results in in the allocation of the target data set on
as many volumes as required, to a maximum of 59
SETSYS VOLCOUNT(ANY)

000-0193-02 SHARE Session 3035 6


DFSMShsm Parmlib
• Indicate which target GDS parameter to pass as a
default during ARECOVER to level 0
Options are:
DEFERRED
ACTIVE
ROLLEDOFF
SOURCE (default)

SETSYS ARECOVERTGTGDS(SOURCE)
000-0193-02 SHARE Session 3035 6
DFSMShsm Parmlib
• Indicate how many ABACKUP or ARECOVER tasks
to run concurrently (up to 64 1.4 or 15 in 1.3 )
SETSYS MAXABARSADDRESSSPACE(15)
• Optionally indicate the utilization percentage for non-
SMS managed volumes during ARECOVER
(default is 80%)
SETSYS ARECOVERPERCENTUTILIZED(080)

000-0193-02 SHARE Session 3035 6


Aggregate Management Class
• SMS must be active
• Create a SMS management class for ABARS
aggregates:
– Create as many as necessary, depending on options
– Use a naming standard for management classes
– Use management class ONLY for ABARS aggregates
– Information on panels 1 - 4 is ignored (use defaults)

000-0193-02 SHARE Session 3035 6


Aggregate Management Class
• In ISMF’s management class creation panel, indicate
the following:
– Number of aggregate versions
• This is the number of active “generations” for each aggregate
using this management class
– Retain Only Version
• If you stop running the aggregate, how long to you want to
keep the most current version (indicate a number then the
indicate Days, Weeks, Months, Years)

000-0193-02 SHARE Session 3035 6


Aggregate Management Class
• In ISMF’s management class creation panel, indicate
the following:
• Retain Extra Versions
• If you stop running this aggregate how long do you want to keep
the extra versions (if versions is more than 1)
• Copy Serialization
• Indicate whether to tolerate an enqueue failure and CONTINUE
processing or to FAIL the backup

000-0193-02 SHARE Session 3035 6


Aggregate Management Class
• In ISMF’s management class creation panel, indicate
the following:
• ABACKUP Copy Technique
– Concurrent Preferred
– Concurrent Required(backup fails if device is not available)
– Standard
• Copy Serialization
• Indicate whether to tolerate an enqueue failure and CONTINUE
processing or to FAIL the backup

000-0193-02 SHARE Session 3035 6


Aggregate Definition
• In ISMF’s option 9, Aggregate Group, define an
aggregate group name
– Use a standard naming convention
– Indicate how many copies of the aggregate to create
concurrently (for example, 1 for offsite, 1 for onsite)
– Indicate the management class name (otherwise the SMS
default management class is used)
– Indicate the output data set name prefix
• Use a unique naming standard that identifies ABARS tapes easily
(will be used in vault pattern description)
• Recommend using the aggregate name as part of the prefix

000-0193-02 SHARE Session 3035 6


Aggregate Definition
• In ISMF’s option 9, Aggregate Group, define an
aggregate group name
– Indicate the selection data set name
– Use a standard naming convention
– Use a PDS, PDSE (OW29117) or sequential data set
Sequential is recommended if the data changes often
– You may indicate up to 5 selection data sets
– Indicate the name of the instruction data set (optional)
• instruction data set not often used
• usually site uses D.R recovery scipt that is written and updated by
all departments involved with recovery.

000-0193-02 SHARE Session 3035 6


Selection Data Sets
• Include/Exclude
– Indicate the data set names either fully qualified or partially
• For generation data sets, you must use the fully qualified name
such as P.SPC.D309877.TRANS.G0001V00 or
P.SPC.D309877.TRANS(0) or a partially qualified base name
with relative generation such as P.SPC.d309877.TRANS.*
• Be careful not to include more generations than you need, many
may be migrated to ML1 or ML2 and cause ABACKUP to run
longer than expected (or necessary) using valueable resources
• Partially qualified GDG names are supported in the INCLUDE
statement but not in EXCLUDE, ALLOCATE,
ALLOCATEEXCLUDE, ACCOMPANY or
ACCOMPANYEXCLUDE.

000-0193-02 SHARE Session 3035 6


Selection Data Sets
• Include/Exclude
– Include (P216.P312522.TUNLDITE.G1270V00
P216.P312522.TUNLDITE.G1271V00
P216.P312522.TUNLDITM.G1270V00
P216.P312522.TUNLDITM.G1271V00
P216.P312522.TUNLDITV.G1268V00
P216.P312522.TUNLDITV.G1269V00
P216.P312522.TUNLDOUT.G1268V00
P216.P312522.TUNLDOUT.G1269V00
P216.P312522.TUNLDPID.G1268V00
P216.P312522.TUNLDPID.G1269V00
P304.P2200008.DW220001.G2914V00
p305.P0050010.TMIFOUT.G4322V00
p305.P0050010.TMIFOUT.G4323V00)
– Exclude (SYS1.**)

000-0193-02 SHARE Session 3035 6


Selection Data Sets
• Allocate/Allocateexclude (optional)
– Allocate (P216.P312522.TUNLDITE.G1270V00
P216.P312522.TUNLDITE.G1271V00
P216.P312522.TUNLDITM.G1270V00
P216.P312522.TUNLDITM.G1271V00
P216.P312522.TUNLDITV.G1268V00
P216.P312522.TUNLDITV.G1269V00
P216.P312522.TUNLDOUT.G1268V00
P216.P312522.TUNLDOUT.G1269V00
P216.P312522.TUNLDPID.G1268V00
P216.P312522.TUNLDPID.G1269V00
P304.P2200008.DW220001.G2914V00
p305.P0050010.TMIFOUT.*)
– AllocateExclude (P304.p0050010.timfout.annual.*)

000-0193-02 SHARE Session 3035 6


Selection Data Sets
• Accompany/Accompanyexclude (optional)
– Accompany (P216.P312522.TUNLDITE.G1270V00
P216.P312522.TUNLDITE.G1271V00
P216.P312522.TUNLDITM.G1270V00
P216.P312522.TUNLDITM.G1271V00
P216.P312522.TUNLDITV.G1268V00
P216.P312522.TUNLDITV.G1269V00
P216.P312522.TUNLDOUT.G1268V00
P216.P312522.TUNLDOUT.G1269V00
P216.P312522.TUNLDPID.G1268V00
P216.P312522.TUNLDPID.G1269V00
P304.P2200008.DW220001.G2914V00
p305.P0050010.TMIFOUT.*)
– AccompanyExclude (P304.p0050010.timfout.annual.*)

000-0193-02 SHARE Session 3035 6


Selection Data Sets
• Allocate and Accompany data sets must not reside in
the include list as well. If they fall under a generic in
the include list, they must be in the EXCLUDE list
Include(pay.first.national.**)
Exclude(pay.first.national.secret.payroll.annual.**)
pay.first.national.secret.payoff.**,
pay.first.national.secret.payoff.annual.**)
Allocate(pay.first.national.secret.payoff.**)
Allocateexclude(pay.first.national.secret.payoff.annual.**)

000-0193-02 SHARE Session 3035 6


Selection Data Sets
• What is a critical data set?
– For rerun recovery, every data set you need to rerun the cycle
at the DR site
– For forward recovery, every data set you need to process the
next cycle at the DR site
• Selection data sets must be updated every time a
change is made to data set names (new, deleted,
changed) for the application that the SDS supports.
– Lack of maintenance is the #2 reason for failure to recover
– #1 is not including critical data sets in the selection data set list

000-0193-02 SHARE Session 3035 6


IBM Redbooks Available
• IBM’s Web site for Redbooks
WWW.REDBOOKS.IBM.COM
• ABARS Books:
• DFSMShsm ABARS and Mainstar
Solutions (SG24-5089-00 April 1998)

000-0193-02 SHARE Session 3035 38


IBM Redbooks on DR
• Fire in the Computer Room - What
Now? (SG24-4211-01 October 96)
• Disaster Recovery Library S/390
Technology Guide (SG24-4210-01
December 96)
• Disaster Recovery Library: Data
Recovery (GG24-3994-00 October 94)
000-0193-02 SHARE Session 3035 38
Biography - Colleen Gordon
• Colleen Gordon is a principal consultant for Mainstar
Software Corporation. Colleen has planned, implemented
and participated in Disaster Recovery exercises for major
corporations throughout the U.S. Colleen co-wrote the IBM
Redbook DFSMShsm ABARS and Mainstar Solutions.
Colleen provides training and assists customers on the
usage and implementation of Mainstar’s disaster recovery
tools. Colleen has over 20 years of experience in
Information Systems and has vast experience with storage
related software and support .

000-0193-02 SHARE Session 3035 38


Biography - Scott Thomas
• Scott Thomas is a Principal Consultant for Mainstar
Software Corporation. Scott is currently assisting
customers in planning, implementing, and exercising
disaster recovery using ABARS and Mainstar’s ABARS
Manager, ASAP, Catalog Base Line and Aggregate Load
Balancer. Scott has extensive and diversified knowledge of
storage and tape related products. Scott has over 10 years
of experience in Information Systems and has supported
many of Chicago’s largest data center operations.

000-0193-02 SHARE Session 3035 38

You might also like