From: | Rafael Martinez <r(dot)m(dot)guerrero(at)usit(dot)uio(dot)no> |
---|---|
To: | Deblauwe Gino <gino(at)useitgroup(dot)com> |
Cc: | "Tiago J(dot) Adami" <adamitj(at)gmail(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: PostgreSQL performance issues |
Date: | 2007-10-23 11:28:48 |
Message-ID: | 471DDAF0.50906@usit.uio.no |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Deblauwe Gino wrote:
> a) I didn't see a reindex in your mail. That's why a backup and a
> restore work and a vacuum doesn't
> http://www.postgresql.org/docs/current/static/sql-reindex.html
> Do this at least daily with that many inserts
>
Hello
I'am sorry to say that this advice does not sound correct.
He is talking about 10,000 transactions and 15,000 new rows per
table/day. This should not require a daily reindex of 12 indexes. It has
to be something else and more information will help to find the problem.
1GB ram sounds maybe to little for a 40GB+ database. It is difficult to
say more without knowing more about the EXPLAIN output for the selects
with problems. It would be good too to know some of the memory values in
postgresql.conf.
I can count with one hand the times that I have had to run a reindex to
fix something in the past years. We have databases with 9millons+
transactions and 2millons+ inserts/updates a day and we have never had
to run a reindex to get a good performance
PS.- RAID-0 for a database is a disaster waiting to happen.
regards
--
Rafael Martinez, <r(dot)m(dot)guerrero(at)usit(dot)uio(dot)no>
Center for Information Technology Services
University of Oslo, Norway
PGP Public Key: http://folk.uio.no/rafael/
From | Date | Subject | |
---|---|---|---|
Next Message | Gokulakannan Somasundaram | 2007-10-23 11:35:20 | Re: [PATCHES] Including Snapshot Info with Indexes |
Previous Message | Heikki Linnakangas | 2007-10-23 11:11:49 | Re: [PATCHES] Including Snapshot Info with Indexes |