[pgpool-general: 3180] Re: Memory leaks

Anselme Rabouan anselme.rabouan at lifecycle-software.com
Fri Sep 19 17:00:25 JST 2014


Hi Muhammad,

I asked to the developer team about that.
They said me that the related client application is using its own pooling
system.
It sends a "SELECT 1;" query to pgpool every minute once the job is over.
As a result, pgpool process never disconnects the backend.
Then, we can see this pgpool child process is never released and we have a
memory leak of about 1GB per 12 hours so I have to kill it manually.

I know we should not use pooling system on the application side anymore
because we now have a pgpool cluster.
Nevertheless, I would know how to set up pgpool in order to be got rid of
such memory leaks.

Thanks,


*Anselme Rabouan | Software Developer | Lifecycle Software Ltd*
T: 0203 0028304

NOTICE
This e-mail (including any attachments) is confidential and is intended for
the above-named person(s). If you are not the intended recipient, please
notify the sender immediately and delete this email from your system and do
not disclose or use for any purpose.

On 18 September 2014 16:56, Muhammad Usama <m.usama at gmail.com> wrote:

> Hi Anselme
>
> I have been looking into this issue, but unable to reproduce the
> memory leak, Can you please share more details on what kind of
> application and queries are causing this issue.
>
> Kind Regards
> Usama
>
>
> On Mon, Sep 15, 2014 at 6:39 PM, Anselme Rabouan
> <anselme.rabouan at lifecycle-software.com> wrote:
> > Hi,
> >
> > We have a pgpool cluster in streaming replication mode, under the latest
> > 3.3.4 version.
> > One of the pgpool child processes has a memory load which is rising about
> > 1GB per 12 hours.
> > I noticed that it is always the same application / query that brings
> about a
> > such behavior.
> > As a result, a "ps" command render something like:
> > ...
> > postgres  5861    0.0  0.0  91496     1260 ?        S    14:18   0:00
> > pgpool: postgres mydb 192.168.1.2(57029) idle
> > postgres  5862    0.0  0.0  91496     1260 ?        S    14:18   0:00
> > pgpool: postgres mydb  192.168.1.2(57020) idle
> > postgres  20640  0.2  0.0  91496     1436 ?        S    Sep11  15:32
> pgpool:
> > postgres mydb  192.168.1.2(56340) idle
> > postgres  27435 0.5 39.9 1874960 1618096 ?    S    Sep14   7:53 pgpool:
> > postgres mydb  192.168.1.2(50879) idle
> > ...
> >
> > And "psql -U postgres -x -c 'select backend_start, query_start,
> > state_change, waiting, state, query from pg_stat_activity where pid
> =2824'"
> > gives:
> > -[ RECORD 1 ]-+------------------------------
> > backend_start | 2014-09-14 15:11:34.13042+01
> > query_start   | 2014-09-15 14:28:34.207835+01
> > state_change  | 2014-09-15 14:28:34.208433+01
> > waiting       | f
> > state         | idle
> > query         | select 1
> >
> > The same, but one minute later gives:
> > -[ RECORD 1 ]-+------------------------------
> > backend_start | 2014-09-14 15:11:34.13042+01
> > query_start   | 2014-09-15 14:29:34.292283+01
> > state_change  | 2014-09-15 14:29:34.292839+01
> > waiting       | f
> > state         | idle
> > query         | select 1
> >
> > and so on...
> >
> > You'll find as attachments, the pgpool configuration file and the log
> part
> > about the suspected pgpool child process.
> >
> > Hope you could help me,
> >
> > Anselme
> >
> >
> > _______________________________________________
> > pgpool-general mailing list
> > pgpool-general at pgpool.net
> > http://www.pgpool.net/mailman/listinfo/pgpool-general
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20140919/bbae7194/attachment.html>


More information about the pgpool-general mailing list