| From: | Andres Freund <andres(at)anarazel(dot)de> |
|---|---|
| To: | Fabien COELHO <coelho(at)cri(dot)ensmp(dot)fr> |
| Cc: | Simon Riggs <simon(at)2ndQuadrant(dot)com>, Heikki Linnakangas <hlinnaka(at)iki(dot)fi>, digoal zhou <digoal(dot)zhou(at)gmail(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Let PostgreSQL's On Schedule checkpoint write buffer smooth spread cycle by tuning IsCheckpointOnSchedule? |
| Date: | 2015-07-03 06:47:22 |
| Message-ID: | 20150703064722.GI30708@awork2.anarazel.de |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 2015-07-03 07:38:15 +0200, Fabien COELHO wrote:
> I've submitted a patch to improve checkpoint write scheduling, including X00
> hours of performance test on various cases. This patch changes significantly
> the load distribution over the whole checkpoint, and AFAICS has been tested
> on rather small cases.
>
> I'm not sure that the power 1.5 is the right one for all cases. For a big
> checkpoint over 30 minutes, it may have, or not, very large and possibly
> unwanted effects. Maybe the 1.5 factor should really be a guc. Well, what I
> really think is that it needs performance measures.
>
> In conclusion, and very egoistically, I would prefer if this patch could
> wait for the checkpoint scheduling patch to be considered, as it would
> basically invalidate the X00 hours of performance tests I ran:-)
These two patches target pretty independent mechanics. If you patch were
significantly influenced by this something would be wrong. It might
decrease the benefit of your patch a mite, but that's not really a
problem.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Noah Misch | 2015-07-03 07:07:21 | Re: copy.c handling for RLS is insecure |
| Previous Message | Michael Paquier | 2015-07-03 06:46:05 | Re: Memory leak fixes for pg_dump, pg_dumpall, initdb and pg_upgrade |