| From: | "Zeugswetter Andreas DAZ SD" <ZeugswetterA(at)spardat(dot)at> |
|---|---|
| To: | "Josh Berkus" <josh(at)agliodbs(dot)com>, <pgsql-hackers(at)postgresql(dot)org> |
| Cc: | "Jan Wieck" <JanWieck(at)Yahoo(dot)com>, "Neil Conway" <neilc(at)samurai(dot)com>, <testperf-general(at)pgfoundry(dot)org> |
| Subject: | Re: [Testperf-general] BufferSync and bgwriter |
| Date: | 2004-12-16 16:24:30 |
| Message-ID: | 46C15C39FEB2C44BA555E356FBCD6FA40184D279@m0114.s-mxs.net |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
> Hmmm, I've not seen this. For example, with people who are having trouble
> with checkpoint spikes on Linux, I've taken to recommending that they call
> sync() (via cron) every 5-10 seconds (thanks, Bruce, for suggestion!).
> Believe it or not, this does help smooth out the spikes and give better
> overall performance in a many-small-writes situation.
The reason is imho that the checkpoint otherwise also syncs all other
writes. These can be writes other backends had to do to replace a buffer.
Linux obviously lacks a mechanism to distribute the IO for cached writes
over time ala bgwriter (or does not do it when already faced with an IO bottleneck).
Andreas
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andrew Dunstan | 2004-12-16 17:44:25 | integer datetimes |
| Previous Message | Tom Lane | 2004-12-16 15:54:30 | Re: join selectivity |