recovery log filling up fast

[smetter]

I'm currently monitoring a tsm server where the recovery log fills up quite fast.
(TSM level 5.5.2.1)
The recovery log is at 13 GB (=the maximum for TSM 5.5)

I'm wondering why the log fills up so fast.
The database backup trigger has to kick in around midnight each day because the log exceeds 80% used.

The normal scheduled database backup is usually around 11 AM (depending on how long the backup stg processes have run), but by midnight, the log is usually at 80% again.

The database is in rollforward mode.

On another tsm environment that I am monitoring, with more or less the same amount of data,
and a recovery log smaller than 13 GB, one db backup per day is enough.

What can I do to adjust things so that only 1 database backup per day is necessary?
Where should I look to find out why the log is filling up so fast?
Could it be because an external scheduler is used that defines heaps of clientactions each day (instead of normzal convential tsm schedules) ?

 

[moon-buddy]

First, find out what is causing the log to fill up. What is running differently on this server versus the other server?

 

[crazytsm]

Check which session is pinning the log by using this command...

1)show logpinned

Cancel the session as this is the one which is increasing the log at a drastic pace.

checkout for any processes running such as backup stg pool and expiration..

the log size is increasing coz there is more data being backed up..

 

[jharris]

If you are not getting any logpinning issues, you could just simply schedule some additional incremental database backups to run throughout the day, perhaps to a sequential FILE device class such that you don't need to waste tapes or hold up tape drives.

Generally a lot of log entries indicates that you may still be suffering from logpinning (as suggested above), but you may have some large file clients with lots of changes going on.