| From: | Josh Berkus <josh(at)agliodbs(dot)com> |
|---|---|
| To: | Stef <svb(at)ucs(dot)co(dot)za> |
| Cc: | pgsql-performance(at)postgresql(dot)org |
| Subject: | Re: Postgres low end processing. |
| Date: | 2003-10-06 18:21:20 |
| Message-ID: | 200310061121.20681.josh@agliodbs.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general pgsql-performance |
Stef,
> => 1. Make sure that the WAL files (pg_xlog) are on a seperate disk from the
> => database files, either through mounting or symlinking.
>
> I'm not sure I understand how this helps?
It gives you better fsync write performance on a low-end disk setup.
Otherwise, the disk is forced to do a hop-back-and-forth between the database
and the xlog, resulting in much slower updates and thus the database tying up
blocks of RAM longer -- particularly if your shared_buffers are set very low,
which they will be.
On RAID setups, this is unnecessary becuase the RAID takes care of disk access
management. But on a low-end, 2-IDE-disk machine, you have to do it.
--
-Josh Berkus
Aglio Database Solutions
San Francisco
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Bruce Momjian | 2003-10-06 18:23:56 | Re: PostgreSQL Beta4 Tag'd and Bundle'd ... |
| Previous Message | Mike Mascari | 2003-10-06 18:16:22 | Re: Generating a SQL Server population routine |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2003-10-06 18:26:10 | Re: reindex/vacuum locking/performance? |
| Previous Message | Neil Conway | 2003-10-06 18:14:29 | Re: reindex/vacuum locking/performance? |