Hi,

Last weekend we experienced the following:

03/20/1999 00:15:36 ANR4552I Full database backup triggered; started as
process 63.
03/20/1999 01:28:36 ANR0985I Process 63 for DATABASE BACKUP running in the
BACKGROUND completed with completion state SUCCESS at 01:28:36.
03/20/1999 01:28:44 ANR4552I Full database backup triggered; started as
process 66.
03/20/1999 02:31:24 ANR0985I Process 66 for DATABASE BACKUP running in the
BACKGROUND completed with completion state SUCCESS at 02:31:24.
03/20/1999 02:31:28 ANR4552I Full database backup triggered; started as
process 67.
03/20/1999 03:15:40 ANR0985I Process 67 for DATABASE BACKUP running in the
BACKGROUND completed with completion state SUCCESS at 03:15:40.
03/20/1999 03:16:05 ANR4552I Full database backup triggered; started as
process 68.
03/20/1999 03:47:37 ANR0985I Process 68 for DATABASE BACKUP running in the
BACKGROUND completed with completion state SUCCESS at 03:47:37.
03/20/1999 03:49:36 ANR4552I Full database backup triggered; started as
process 69.
...
...

Time after time, a database backup was triggered almost immediately after
the last one completed. This happened 12 times before the system (probably)
ran out of scratch tapes:

03/20/1999 09:23:20 ANR4552I Full database backup triggered; started as
process 91.
03/20/1999 09:25:35 ANR8779E Unable to open drive /dev/rmt2, error
number=47.

03/20/1999 09:25:44 ANR1404W Scratch volume mount request denied - mount
failed.

03/20/1999 09:25:44 ANR4578E Database backup/restore terminated - required
volume was not mounted.

03/20/1999 09:25:44 ANR0985I Process 91 for DATABASE BACKUP running in the
BACKGROUND completed with completion state FAILURE at 09:25:44.

03/20/1999 09:25:46 ANR4560I Triggered database backup will be retried in 60
seconds.

After two more retries, the server went into some kind of coma, that could
only be resolved by expanding the recovery log before restarting the server
(killing and restarting the server without doing this didn't even produce
any messages in the activity log).

Why does this happen? It looks to me like the recovery log isn't emptied
after the database backup. Some additional info:

*       ADSM 3.1.1.5 on AIX 4.2.1 (or rather: a 3466 model C20 Network
Storage Manager with currently installed 3466 Software Level EC: 3.1.1.5)
with a 3575 tape library
*       Log mode is set to roll-forward
*       Full database backup schedules twice a day
*       Unscheduled database backups hardly ever happen

Any help would be greatly appreciated.

Martijn Grendelman.