On Wed, Oct 25, 2006 at 03:54:17PM -0400, Mike Goldner wrote:
> Finally, and most important is the blocking. The vacuum duration
> reported in the log converts to about 170 minutes. I can track
> backwards in the log and the only messages prior to the 6:52am
> completion of the vacuum end at 3:57am (almost exactly 170 minutes
> prior).
>
> So, all indications point to postgres blocking all access during the
> entire vacuum.
The vacuum command you posted doesn't block. I can think of two
other explanations:
1. You just didn't have any activity then. Don't throw away
this possibility without evidence: I can't count the number of blind
alleys I've been down because someone insisted that "never happens".
2. You're completely pegged on I/O. Vacuum will make this
worse, and maybe therefore no transactions get through. You can
fiddle with the vacuum settings to get them to back off a little and
let some other transactions through.
That said, your essential problem is that one table. Vacuum it more
often -- from the look of the churn on it, I'd just put a job on it
that runs all the time and sleeps for a few seconds in between -- and
you shouldn't have this problem. But you'll need to VACUUM FULL or
dump and reload first.
A
--
Andrew Sullivan | ajs@crankycanuck.ca
Windows is a platform without soap, where rats run around
in open sewers.
--Daniel Eran
