David,
Our experience of this was that a client session had stuck in RecVW
over a couple of nights. The effect of this was that that part of the
log after the uncommitted client transaction could not be cleared
( the log in rollforward mode is sequential ). Thus one dark and stormy
night our server got into a cycle of:
1. log fills to trigger incremental backup
2. Backup does not free log below trigger
3. Log triggers another incremental backup
Eventually a full backup will be triggered but still this will not help.
In our case - as this happened overnight - the pool of scratch tapes was
exhausted, the log filled and crashed the server.
We recovered by extending the log, backing up the DB (this didn't reduce
the log to 0% though) , changing the log to normal mode to reduce it again,
then server restart again and change to rollforward mode with a consequent
full db backup.
So, all in all a pain, very time-consuming and, potentially, damaging.
I would advise checking your actlogs and client sessions.
Of course, the fact that the server behaves in such a way is unacceptable
and logically simple to fix. If your message does relate to this
you might like to raise it with Tivoli. I know others have and more
pressure might persuade them to address this problem.
Regards
Ian Smith
--------------------------------------------------------------------------------
Ian Smith email: ian.smith AT oucs.ox.ac
DOT uk
Ian Smith email: ian.smith AT oucs.ox.ac
DOT uk
Oxford University Computing Services
--------------------------------------------------------------------------------
>
>
>Hello,
>
>I am running 4.1.2.0 and my problem is:
>
>ANR4556W Attention: the database backup operation did not
> free sufficient recovery log space to lower
>utilization
> below the database backup trigger. The recovery
>log size
> may need to be increased.
>
>I have run a databasebackup, tryied to switch the recoverylog from
>rollforward to normal mode without luck, can anybody help me?
>
>Regards,
>
>David
|
