Hi all!
First of all thank you for this awesome database! We are successfully using PG on top of Linux
in our projects for 2 years so far and are very happy with the performance and observability
features of the database.
Recently we deployed PG 9.0.5-x64 on Windows Server 2008 R2 and noticed very strange behavior:
at some time after firing up the DB it starts logging many 'pgstat wait timeout' messages.
We monitor the IO, and it never goes higher than 10MB/s, whereas the total throughput of DB disks is ~200MB/s.
I wouldn't bother, but the performance degrades soon after the message starts showing up.
After some digging around I found that the reason is that $PGDATA/pg_stat_tmp/pgstat.stat stops updating.
The workaround for me was to SIGHUP the DB -- pgstat.stat starts updating and the warning stops.
Is this expected behavior?
If this is a bug what information will be helpful in this particular situation to include in the bug-report?
Thanks!