DB2 OS/390 Disaster Recovery questions

[login to unmask email]

DB2 OS/390 Disaster Recovery questions
We ironically completed a Disaster Recovery exercise last week, recovering
subsystems on both DB2 V3 and DB2 V4.
We coincidently had problems with SYSLGRNX. We noticed that while the V3
recoveries applied logs as expected, the V4 recoveries did not.

The V4 recovery of SYSLGRNX left SYSLGRNX and its indexes in a state of RW,LPL.
Subsequent application recoveries do not apply logs at all. The REPORT utility
indicated SYSLGRNX was missing log ranges. Checking the V4 manual revealed we
had omitted the recovery of the SYSLGRNX indexes introduced in V4!

This may not be your problem, if the catalogue applied logs correctly. If your
image copies are on tape, check that a reply of NO was not issued to the request
for the tape containing the image copy or that an I/O error occurred, if not
hopefully a copy of the system log will give you a clue.

Message DSNU513I csect-name RECOVER UTILITY LOG APPLY RANGE IS xxx TO yyy is
issued by the RECOVER utility
when archive logs are applied.

Robin Leske
Database Administrator
Ph: +61 8 8233 1308
Email: [login to unmask email]
---------------------- Forwarded by Robin LESKE/SGIOGRP on 06/10/99 06:08 pm

[login to unmask email] on 02/10/99 12:29:31 am

Please respond to [login to unmask email]

To: Multiple recipients of list db2-l <[login to unmask email]>
cc: (bcc: Robin LESKE/SGIOGRP)
Subject: DB2 OS/390 Disaster Recovery questions


We recently completed a Disaster Recovery (DR) exercise and would like
to address 2 situations which occurred.

Our systems personnel recovered all DB2 volumes to 19 Sep. The
disaster occurred at 0600 hrs on 24 Sep. We had an image copy
of all tablespaces including catalog and directory from 0200 hrs on 24
september and archive logs through 0419 hrs on 24 sept. I.E. We
could recover at the DR site up to 0419 hrs on 24 Sep.

(1) During the recovery of the catalog and directory (we followed the
order specified in UTILITY guide), all tablespaces used the latest
image copy (24 sep) & the latest ARCHLOG tape (from 24 Sep). However,
SYSLRGNX did not use an image copy and read ALL ARCHLOGS from 19-24
Sept.Why did SYSLGRNX do this????Shouldn't SYSLGRNX use the latest
image copy and then read the latest ARCHLOG tape like the other
tablespaces?? Please explain.

(2) For the RECOVERY utility, is there any way to display whether or
not ARCHLOG tapes are being used 'during' RECOVERY execution? I know
the CONSOLE LOG and the REPORT utility show this information, but I
would like to see this info in the RECOVERY JCL output as the RECOVERY
utility is being executed.

Thank you for your time.