In response to "Pavel Rotek" <pavel.rotek@gmail.com>:
> 2008/3/7, Tom Lane <tgl@sss.pgh.pa.us>:
[snip]
> > If you have to do it that way, you'll need very frequent vacuums on this
> > table (not vacuum full, as noted already) to keep the toast space from
> > bloating too much. And make sure you've got max_fsm_pages set high
> > enough.
>
> i'll set max_fsm_pages to 1 000 000. It should be enough and set
> autovacuum_naptime to 10 minutes. May it be?
No. Keep naptime at 1 minute. If it comes around and there's nothing
to do, the overhead is minimal. If you set the naptime too high, it might
have too much to do on the next cycle and then it'll bog things down.
Also, it only checks 1 database per cycle, so setting it to 10 minutes
means a _minimum_ of 40 minutes between checks (because you have a template0,
template1, postgres, and your database minimum)
Also, keep an eye on your database bloat to ensure the various
autovacuum_*_scale_factor and related settings are appropriate.
It's been found that these are often not aggressive enough for
good maintenance. If you see bloat even with autovacuum running,
reduce those values.
Personally, I'd recommend running a MRTG graph that graphs the size
of this table so you can easily watch to see if your config tweaks
are getting the job done or not. And remember that _some_ bloat is
expected and normal for operation.
--
Bill Moran
Collaborative Fusion Inc.
http://people.collaborativefusion.com/~wmoran/
wmoran@collaborativefusion.com
Phone: 412-422-3463x4023
