What is causing so many mboxlist DB Transaction Log Files ?
(Doc ID 1494565.1)
Last updated on JUNE 11, 2024
Applies to:
Oracle Communications Messaging Server - Version 5.1.0 to 7.0.4 [Release 5.1 to 7.0.0] Information in this document applies to any platform.
Goal
As compared to other similarly configured and loaded message store systems, this one system is processing transaction log files at a much higher rate. What is it doing ?
For example, in the mboxlist.ls captured by dbhang, we see:
% cat mboxlist.ls
total 1668821
drwx------ 4 mailsrv mail 19 Sep 23 19:30 .
drwx------ 6 mailsrv mail 6 Oct 18 2011 ..
-rw------- 1 mailsrv mail 0 Oct 29 2011 .loglock
-rw------- 1 mailsrv mail 409600 Sep 23 08:31 annotate.db
-rw------- 1 mailsrv mail 333160448 Sep 23 19:29 folder.db
-rw------- 1 mailsrv mail 10485760 Sep 23 19:29 log.0000151979
-rw------- 1 mailsrv mail 10485760 Sep 23 19:29 log.0000151980
-rw------- 1 mailsrv mail 10485760 Sep 23 19:29 log.0000151981
-rw------- 1 mailsrv mail 10485760 Sep 23 19:30 log.0000151982
-rw------- 1 mailsrv mail 10485760 Sep 23 19:30 log.0000151983
-rw------- 1 mailsrv mail 10485760 Sep 23 19:30 log.0000151984
-rw------- 1 mailsrv mail 10485760 Sep 23 19:30 log.0000151985
-rw------- 1 mailsrv mail 10485760 Sep 23 19:30 log.0000151986
drwx------ 2 mailsrv mail 2 Sep 23 19:29 logarchive
-rw------- 1 mailsrv mail 8192 Oct 18 2011 lright.db
drwx------ 2 mailsrv mail 15 Sep 23 19:19 mqueue
-rw------- 1 mailsrv mail 321916928 Sep 23 19:29 peruser.db
-rw------- 1 mailsrv mail 41825280 Sep 23 19:30 quota.db
-rw------- 1 mailsrv mail 78249984 Sep 23 19:29 subscr.db
There are several transaction log files created at the same time. The above was collected at 19:30:29. So in the last 90 seconds, 8 transaction log files were created.
The log files are processed and removed by stored in a timely fashion. The problem is not that they are accumulating. In fact, no performance problem is being noticed. There is just the question of why the system is creating so many transaction log files so quickly?
Also, based on the snapshots.ls, we see the numbers of the transaction log files at the time of the most recent snapshot:
<server-root>/data/store/dbdata/snapshots/005:
total 1536161
drwxr-xr-x 8 mailsrv mail 9 Oct 29 2011 ..
-rw------- 1 mailsrv mail 78217216 Sep 16 02:00 subscr.db
-rw------- 1 mailsrv mail 8192 Sep 16 02:00 lright.db
-rw------- 1 mailsrv mail 393216 Sep 16 02:00 annotate.db
-rw------- 1 mailsrv mail 0 Sep 16 02:00 .verified
-rw------- 1 mailsrv mail 41825280 Sep 16 02:00 quota.db
-rw------- 1 mailsrv mail 333004800 Sep 16 02:00 folder.db
-rw------- 1 mailsrv mail 321732608 Sep 16 02:00 peruser.db
-rw------- 1 mailsrv mail 10485760 Sep 16 02:00 log.0000136769
-rw------- 1 mailsrv mail 0 Sep 16 02:05 .snaptime
drwx------ 2 mailsrv mail 11 Sep 16 02:05 .
Note: the most recent snapshot being several days old was a separate problem.
From this, you can see the average transaction log file rate over several days has been 2,000 per day -- about 85 per hour -- sometimes spiking to several per minute. On other similarly configured and loaded systems, it is only a few per hour.
Solution
To view full details, sign in with your My Oracle Support account.
Don't have a My Oracle Support account? Click to get started!