I have a similar problem with one of our larger servers.
The server holds a 50 GB database.

The latest crash occured this morning. The recovery log exceeded the
limit of 75%. The triggered database backup did not free up sufficient space,
db backup started again and so on. Normally we stop the server before we
run out of recovery log space. This morning, 5:00 am, nobody was present to do
it.
So the server crashed. After the restart everything was fine.

The server holds a 50 GB database. The recovery log is 4 GB.
At the time of the first crash the trigger was set to 50%.  Following the
recommendation of the error message text we increased the trigger to 75%. All
the same the db looping continued.
I have an open record for this problem.

Werner


Joel Fuhrman wrote:

> AIX 4.2.1
> ADSM Server level 3.1.2.13
>
> I have defined my dbbackuptrigger to be 50%.  The max log shows 57%.  When
> the trigger was activated, it failed to reduce the log to less than 50%.  Is
> this a bug?
>
> As the backups continued, it triggered additional dbbackups. Eventually, the
> activity log is flooded with these messages:
> 1999-02-09 03:06:40      ANR9999D logread.c(179): Attempt to read LSN
>                           18717.193.3351 below trunc LSN 18768.171.991.
>
> Finally, the system paging space filled and processes were killed by AIX
> which had the effect of killing ADSM.  Has any one else seen this behavior?
>
> ADSM is the only thing running on this system.  The paging space is normally
> only 22% used.

--
Werner Baur                    Tel.:++49-89-28928781
Werner Baur                    Tel.:++49-89-28928781
Leibniz-Rechenzentrum           Fax:++49-89-2809460
Barer Str. 21                mailto:baur AT lrz DOT de
D-80333 Muenchen, Germany      http://www.lrz.de