Backup process for Mainframe Disaster
Recovery
(02/17/10)
Prod LPAR
00) Clean the tape drives
01) Submit INTAPES
and print the output.
02) Pull and set aside
the tapes listed in INTAPES job.
03) Submit GENTAPES
04) When GENTAP1
finishes, set aside the tapes from INTAPES and return them to the Live rack.
05) Set aside the
output from GENTAP1
06) Submit
FLASHASK
----- Submit this several times if necessary.
Run/view until
flash copies are done.
Look in Sysout for each address:
Pertinent
data:
Value under 'DEVN' is address being
backed up.
Value under 'ACT' is status of copy
process
0 = Copy
complete
1 = Copy still running
Examples:
... DEVN SSID LSS CCA CU
SERIAL ACT ....etc...
... 050B 1102 02
0B 2105 000000023937 0 ....etc...
means address 050B Copy
complete
... DEVN SSID LSS CCA CU
SERIAL ACT ....etc...
... 0900 1100 00
0B 2105 000000023937 1 ....etc...
means address 0900 Copy
still running
When ACT status
= 0 for ALL
addresses, Flashcopy is done.
07) Submit TAPEOFF
Test LPAR
08) Submit TAPEON
Submit
ONLINE
09) Submit
DUMPTEST
Release jobs DTESTxx per available tape drives
10) Verify all jobs
ran successfully.
If I/O errors require a pack be re-dumped:
Put defective tape(s) aside.
Customize DUMP0001, submit it, CANCEL
11) Submit
LISTTEST
12) Select SYSOUT
for LISTTEST
PRINT ODSN
'RLS0.SUNGARD.TEST'
PRINT
PRINT CLOSE
13) Submit
TAPEOFF
Submit
OFFLINE
14) Submit
REINITPX
15) Submit SYS1.RMM.SYSJOBS(RMMDAILY)
16) Set aside all
output tapes
Prod LPAR
17) Submit TAPEON and clean the tape drives
Stop here - wait for HSM migration at 1pm
DFSMSHSM will have
written tapes at 1pm.
The tape numbers can be
determined by logging on TSO at 1:15pm
S:PRE DFSMSHSM;DA
? to
left of job, to ‘explode’ the Sysout.
S to left of JESMSGLG, M on command line and PF8 to go
to end of Sysout.
F FULL
You will see (for example)
ARC0421I MIGRATION VOLUME 006939 IS NOW MARKED FULL
ARC0443I MIGRATION VOLUME 006939 WITH ALTERNATE 005022
ARC0443I (CONT.) IS NOW MARKED FULL
where 006939 is volser of Primary
tape, 005022 is volser of duplex tape
Volumes 006939 and 005022 would
go into our ‘live’ tape rack.
You may now resume the Prod LPAR process
18) Submit
DUMPPROD
Release jobs DPRODxx per available tape drives
19) Submit
LISTPROD
If I/O errors require a pack be re-dumped:
Put defective tape(s) aside.
Customize DUMP0001, submit it, CANCEL
20) Select SYSOUT for
LISTPROD
PRINT ODSN
'RLS0.SUNGARD.PROD'
PRINT
PRINT CLOSE
21) Submit
REXXRRRR
SAVE OUTPUT
TAPE!
Save listings
from printer
22) Print two copies
of listing from Step 16
Via SDSF
PRE
REXXRRRR
H
PRINT ODSN
'RLS0.PRINT.PRINT(DAILY)'
PRINT
PRINT
CLOSE
/
/
23) Save electronic
copy of listing:
Submit PMHH.REXX.JCL(BKUPLIST)
and verify FTP successful
24) Verify tapes against listings
Pack tapes
and one listing for
File second listing in book.
25) Create Iron
Mountain forms via automated procedure
26) FTP Restore Procs and JCL to a PC and copy to a thumbdrive which goes to IM with the tapes.
Handling Tape Errors
If the system gets I/O errors on an output tape, it may offer to ‘swap’ to a different drive. Always say ‘NO’ to this offer. The particular jobstep will end with an error code and the job will Resume with the next step.
Examine the job to see what pack was being dumped when the error occurred.
For example, if the pack being unsuccessfully dumped were POL003:
Edit SYS1.SUNGARD.FLASH.BACKUP.TEMP(DUMP0001):
//DUMP0001 JOB 1,RLS,MSGLEVEL=(1,1),MSGCLASS=H,TYPRUN=HOLD
//* CHANGE ?????? TO THE VOLSER OF THE PACK TO DUMP
//* SUBMIT JOB, THEN CANCEL TO NOT SAVE THE CHANGE
//D?????? EXEC DRBKPROC,V=??????,RET=2,FREQ=DLY
Enter command C ?????? POLOO3 ALL
You will now see:
//DUMP0001 JOB 1,RLS,MSGLEVEL=(1,1),MSGCLASS=H,TYPRUN=HOLD
//* CHANGE POL003 TO THE VOLSER OF THE PACK TO DUMP
//* SUBMIT JOB, THEN CANCEL TO NOT SAVE THE CHANGE
//DPOL003 EXEC DRBKPROC,V=POL003,RET=2,FREQ=DLY
Enter SUBMIT
Enter CANCEL
F3
Go to SDSF and release DUMP0001 (when a tape drive is free).
The defective tape will have been ejected but will not show up in the final tape list – it will be an ‘extra’ and should be put on my desk.