The Vacuum's don't take too long, 10 minutes at most.  I can tell from ps
-ef | grep and top that its the selects/inserts/updates from the postgres
related to our app that take all that time up.  If we rerun initdb and
reload the data, it works great for about two days, then goes bad again.

We are in the process of trying out 7.4.2 right now, just waiting on the
reload of pg_dump.


> Fabio,
> > We run VACUUM ANALYZE after we remove about 1000 rows every hour on the
> > halh hour.  Our max_fsm_pages is set to 10000
> Have you checked how long these vacuums take?   If they are starting to
> overlap, that would explain your high CPU usage and poor performance.    You
> might want to consider raising FSM_pages and vacuuming less frequently.
> --
> -Josh Berkus
>  Aglio Database Solutions
>  San Francisco
> ---------------------------(end of broadcast)---------------------------
> TIP 2: you can get off all lists at once with the unregister command
>     (send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])

---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
      joining column's datatypes do not match

Reply via email to