From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | revoohc(at)gmail(dot)com |
Cc: | pgsql-admin(at)postgresql(dot)org |
Subject: | Re: Question on vacuuming |
Date: | 2005-09-13 17:46:22 |
Message-ID: | 25350.1126633582@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
Chris Hoover <revoohc(at)gmail(dot)com> writes:
> there is 1 exception. Everytime there is work to do, it seems to spike our
> load from < 1 to a load of 3+ for a short time. This tends to cause
> significant slowdowns for our customer if they happen to do anything that is
> causing the database to access the drives. It does not seem to matter how
> much work there is to do. Cleaning up 1 page/tuple or 15000 page/tuples
> tends to cause this spike.
No surprise, because vacuum has to read every page of the table to see
if there is anything there to clean up.
> We are running 7.3.4 on Dell 2650 connected to a Dell powervault running
> raid5.
If you were running something a little newer (8.0) you could use the
vacuum_cost_delay knobs to reduce vacuum's impact on I/O load.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Marcin Giedz | 2005-09-13 18:03:52 | Re: plperl again but different problem |
Previous Message | Marcin Giedz | 2005-09-13 17:44:28 | Re: plperl again but different problem |