| From: | Stefan Kaltenbrunner <stefan(at)kaltenbrunner(dot)cc> |
|---|---|
| To: | robert(at)webtent(dot)com |
| Cc: | PostgreSQL <pgsql-general(at)postgresql(dot)org> |
| Subject: | Re: Getting all tables into memory |
| Date: | 2008-01-24 20:17:14 |
| Message-ID: | 4798F24A.8080505@kaltenbrunner.cc |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
Robert Fitzpatrick wrote:
> I have a couple of servers running Postfix with amavisd-maia+SA+clamav
> on FreeBSD 6.2 which use a central db server running PgSQL 8.2.4. My
> issue is the bayes database causing SA TIMED OUT in the logs and want to
> make sure I am getting everything into memory. The disk activity is high
> on the db server, this is the average systat status...
The SA SQL-bayes implementation is one prime example of a real life
application that can benefit from HOT because it is basically updating a
fairly limited set of non-indexes columns at an insane rate.
I have seen real live installations that could barly keep up with bloat
even on a tight 3min vacuum cycle and the 8.3B4 test instance I have
here can take at least 4 times the load than 8.1 could using that kind
of workload.
Stefan
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Dave Page | 2008-01-24 20:19:41 | Re: Installing PostGreSQL - 2 |
| Previous Message | Wes | 2008-01-24 19:35:00 | Re: REINDEX on large DB vs. DROP INDEX/CREATE INDEX |