Re: Strange behaviour under heavy load

From: "Dmitry Koterov" <dmitry(at)koterov(dot)ru>
To: "Oleg Bartunov" <oleg(at)sai(dot)msu(dot)su>
Cc: "Postgres General" <pgsql-general(at)postgresql(dot)org>
Subject: Re: Strange behaviour under heavy load
Date: 2007-03-28 10:12:43
Message-ID: d7df81620703280312n34d4e2besdb806c3f488a402@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Disc write activity timeline looks like the following:

CHECKPOINT ->
Large disc write (30-40M/s), 1 second ->
4-5 seconds: no read and write activity at all, but server works slow ->
2-3 seconds: disc write activity about 4-5 M/s ->
restore normal speed

On 3/28/07, Dmitry Koterov <dmitry(at)koterov(dot)ru> wrote:
>
> New information.
>
> When I run CHECKPOINT manually in psql console, effect is fully
> reproduced.
> So - it is a checkpoint slowdown.
>
> The quesion is - how to make CHECKPOINT work faster?
>
> On 3/27/07, Dmitry Koterov <dmitry(at)koterov(dot)ru> wrote:
> >
> > How to tune them?
> >
> > Now I have:
> >
> > checkpoint_segments = 5 # in logfile segments, min 1, 16MB each
> > checkpoint_timeout = 900 # range 30-3600, in seconds
> > checkpoint_warning = 30 # in seconds, 0 is off
> >
> > No checkpoint warnings in pgsql logs.
> >
> >
> > On 3/27/07, Oleg Bartunov < oleg(at)sai(dot)msu(dot)su> wrote:
> > >
> > > hmm,
> > >
> > > looks like checkpoint ?
> > > Check checkpoint settings in postgresql.conf.
> > >
> > >
> > > Oleg
> > > On Tue, 27 Mar 2007, Dmitry Koterov wrote:
> > >
> > > > Hello.
> > > >
> > > > I have found that sometimes heavy loaded PostgreSQL begins to run
> > > all the
> > > > queries slower than usual, sometimes - 5 and more times slower. I
> > > cannot
> > > > reprocude that, but symptoms are the following: queries work very
> > > fast for
> > > > 5-10 minutes, and after that - significant slowdown (every query,
> > > even a
> > > > simple one, works 5-10 and ever more times slower), disk write
> > > activity
> > > > grows too (but not everytime - I don't know exactli if there is a
> > > direct
> > > > correlation). 2-3 seconds, then - performance restores back to
> > > normal.
> > > >
> > > > Autovacuum is turned off.
> > > >
> > > > Machine has 2 processors, huge memory, fast SCSI disks.
> > > >
> > > > I understand that there is too less information. Please advice what
> > > to
> > > > monitor for better problem discovering.
> > > >
> > >
> > > Regards,
> > > Oleg
> > > _____________________________________________________________
> > > Oleg Bartunov, Research Scientist, Head of AstroNet ( www.astronet.ru
> > > ),
> > > Sternberg Astronomical Institute, Moscow University, Russia
> > > Internet: oleg(at)sai(dot)msu(dot)su, http://www.sai.msu.su/~megera/<http://www.sai.msu.su/%7Emegera/>
> > > phone: +007(495)939-16-83, +007(495)939-23-83
> > >
> > > ---------------------------(end of
> > > broadcast)---------------------------
> > > TIP 1: if posting/reading through Usenet, please send an appropriate
> > > subscribe-nomail command to majordomo(at)postgresql(dot)org so that
> > > your
> > > message can get through to the mailing list cleanly
> > >
> >
> >
>

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Albert 2007-03-28 12:02:57 Paranthesis and Driver JDBC
Previous Message Ragnar 2007-03-28 10:11:42 Re: Query that does not use indexes