On 19/11/2009 1:23 AM, Scott Felt wrote:
>
> Hello. I have been having an issue with a database. The logs consistently
> show this:
[snip]
> Prior to this, the logs show nothing other than AutoVacWorkerMain
> activities. This database is installed on a Windows OS. The version of
> PostgreSQL is 8.4.1 (which I was hoping had a fix since I saw references to
> a bug fix for a shared memory problem). Does anybody have any suggestions
> on how to determine what causes these shutdowns? Should I be considering
> dialing up the shared_buffers size in the config file? Currently it is set
> at 32mb.
The shared memory issue - "cannot reattach to shared memory" - doesn't
appear to be what's happening here.
It'd really help to know why one of the server processes crashed in the
first place:
2009-11-17 16:03:55 ESTLOG: 00000: server process (PID 9644) exited
with exit code 128
Do you know what process this was? Was it autovacuum? A query backend?
Is there anything in particular you're doing when it crashes? Or does it
just crash randomly? Even when left idle? If so, does it sometimes crash
almost immediately, or does it always take a while before it crashes?
ie: is the delay between startup and crash predictable?
Do you have any antivirus software installed? Even if it is deactivated?
--
Craig Ringer