| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | kris pal <kristhegambler(at)yahoo(dot)com> |
| Cc: | pgsql-admin(at)postgresql(dot)org |
| Subject: | Re: regression database |
| Date: | 2004-08-27 03:29:21 |
| Message-ID: | 5707.1093577361@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-admin |
kris pal <kristhegambler(at)yahoo(dot)com> writes:
> Actually I just noticed something. Previously for some reason, I
> thought the data directory corresponded to 'regression'. But I double
> checked it now with oid2name and it corresponds to our main database
> say "dds".
Ah, that makes more sense.
> But the issue still remains. In the sense that pg_attribute etc are
> still so huge and I have no idea why.
Probably because you've created and deleted a lot of tables (does your
app use temp tables a lot?). If you aren't good about vacuuming the
system catalogs then you're going to get table bloat. If you're using
something pre-7.4 then you may get index bloat even if you *are*
vacuuming regularly :-(.
My suggestions are to update to 7.4, if you're not there already
(if you are, you may need to do VACUUM FULL and REINDEX to get rid
of the bloat); then adopt a stricter regime of routine vacuuming.
And check your FSM settings.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andrew Farmer | 2004-08-27 04:03:59 | Re: 7.3.4 Table corruption |
| Previous Message | kris pal | 2004-08-26 21:15:32 | Re: regression database |