Yes, because all these messages really mean is that autovacuum woke up,
and looked at database xx, it may or may not have decided to do anything
(analyze or vacuum) and judging by the fact that these messages are
almost exactly 1 minute apart, I would say that autovacuum never felt
the need to take action during this log segment.
One of the deficiencies in autovacuum in 8.1 is that it's not easy to
know when autovacuum actually did take action. This is somewhat
improved in 8.2.
Michael Goldner wrote:
I notice in my postgres logfile that autovacuum is running every minute.
The server had no overnight activity, yet this has been going on for days.
Overall performance of the database seems to be slow, and I assume that it
is related.
I'm running 8.1.4 on Redhat EL4.3.
Is this normal for autovacuum?
Here is the tail of the log (even though it doesn't say much):
[4836---2007-02-13 08:28:28.194 EST]LOG: autovacuum: processing database
"xx"
[4859---2007-02-13 08:29:28.200 EST]LOG: autovacuum: processing database
"xx"
[4863---2007-02-13 08:30:28.196 EST]LOG: autovacuum: processing database
"xx"
[4865---2007-02-13 08:31:28.191 EST]LOG: autovacuum: processing database
"xx"
[4870---2007-02-13 08:32:28.201 EST]LOG: autovacuum: processing database
"xx"
[4916---2007-02-13 08:34:23.555 EST]LOG: autovacuum: processing database
"xx"
[4921---2007-02-13 08:35:23.550 EST]LOG: autovacuum: processing database
"xx"
[4923---2007-02-13 08:36:23.555 EST]LOG: autovacuum: processing database
"xx"
[4959---2007-02-13 08:37:27.169 EST]LOG: autovacuum: processing database
"xx"
[4965---2007-02-13 08:38:27.165 EST]LOG: autovacuum: processing database
"xx"
[5057---2007-02-13 08:39:27.178 EST]LOG: autovacuum: processing database
"xx"
Thanks,
Mike
