On Mon, 15 Oct 2001 14:24:29 +0100, it was written:
>In normal daily processing the recovery log usually gets up to 1.6 gb.  Last
>Thursday night however the log filled up and I had to add more space with
>the usual dsmfmt and dsmserv utilities.  My daily reports show no unusual
>backup/archive activity other than the database buffer pool was overcommited
>by 12 kilobytes.
>
>The amount of space used implies something like 1.5 million files additional
>files wished to be backed up.  I can see no trace of these files anywhere.
>
>Does anyone know of any bugs in the above environment which could cause this
>to happen?

This problem has been discussed numerous times on this list in the
last year, and thus also appears in the adsm.org archives.

Log entries are committed and deleted from the log when a database
backup is done. Occasionally, if other processes are running when the
log commits are running (shortly after the completion of the db
backup), those other processes will "step on" the log entries,
preventing commitment from completing and thus not clearning the log.
Thus the log could fill to capacity and bring the TSM server down.

--
Mark Stapleton (stapleton AT berbee DOT com)
Mark Stapleton (stapleton AT berbee DOT com)