For the past few days, my daily maintenance has been interrupted by the
database backup. When it's invoked, it never happens, and the maintenance
script of course dosn't finish. I can't do it manually, either. While it's in
that state, I can't get answers on any sort of storage query - select on
volumes, q stg, q drm...
If I halt and restart the server twice, I can get it - our I could, until
today. Today I've done two restarts with no luck, even with client sessions
disabled, so I'm trying switching logmode to normal and back to rollforward to
get the triggered backup.
There's nothing of use in the activity log. I've got scratch, I've got all
drives, paths, and libraries online (working fine for bastg). I am finding
nothing relevant in the available APARs. I reckon I'll try enabling all the
"DB" trace classes and taking a run at deciphering them.
While I was typing, the server process finished restarting for the third time,
and switching logmode did trigger the backup. So I'm OK for today, but I don't
want to have to dial in every day of any time off to work around this. I wish
desperately that my company considered their absolute commitment to 100% data
protection to include keeping up support on the central piece of that function.
Has anybody seen this sort of thing? Anybody got a better workaround? Maybe
an APAR I missed? I'm campaigning for permission to update to 5.5.4, but I
don't see anything to suggest that a problem like this is fixed by it.
Another detail: This sort of thing has happened the last few times I've
generated the month-end backupsets. Previously, one halt, and start the next
backupset as soon as the dbb is done so it's finished by the next day, and
everything fihishes fine. This time I stopped my backupset generations 5 days
ago and the problem has remained.
73,
Tim Conway
JBS USA | 1770 Promontory Ci | Greeley, CO 80634| USA
Direct: 970-506-7998 | Fax: 970.336.6195
email: Timothy.Conway AT jbssa DOT com<blocked::mailto:Timothy.Conway AT jbssa
DOT com>
JBS Server Team
|