[pgpool-hackers: 1299] Re: pgpool zombie with wathdog 3.5

Muhammad Usama m.usama at gmail.com
Tue Jan 5 04:37:57 JST 2016


On Mon, Jan 4, 2016 at 10:16 AM, Yugo Nagata <nagata at sraoss.co.jp> wrote:

> Hi Usama,
>
> On Tue, 29 Dec 2015 03:32:00 +0500
> Muhammad Usama <m.usama at gmail.com> wrote:
>
> > Hi Yugo
> >
> > Many thanks for the patch. Although the patch takes care of zombie
> > processes, but I think blocking the watchdog process at time of
> escalation
> > and de-escalation might lead to some un-expected results, because at that
> > time when the watchdog process will be waiting for the
> > escalation/de-escalation process to finish it would not be replying to
> > other nodes.
> > So can you please have a look at the attached path which uses the
> SIGCHILD
> > to handle the zombie process problem
>
> I have looked at and tested this and I think this is good to be committed.
>

I have committed the patch to master branch
http://git.postgresql.org/gitweb?p=pgpool2.git;a=commitdiff;h=106681f4bf6a5ecb230716c33918a0d76c4cdcfc


Best regards
Muhammad Usama


> >
> > Thanks
> > Best regards
> > Muhammad Usama
> >
> > On Mon, Dec 28, 2015 at 2:11 PM, Yugo Nagata <nagata at sraoss.co.jp>
> wrote:
> >
> > > Hi Usama,
> > >
> > > I write the patch to fix it.
> > >
> > > In current, escalation process is forked but not waited by watchdog
> > > process. Fix is to wait the escalation process by waitpid. That is,
> > > watchdog is blocked while waiting the escalation finishing.
> > >
> > > Any comment?
> > >
> > > Regards,
> > >
> > > On Mon, 21 Dec 2015 16:46:33 +0900
> > > Yugo Nagata <nagata at sraoss.co.jp> wrote:
> > >
> > > > Usama,
> > > >
> > > > A zombie process appeared after pgpool started with watchdog.
> > > >
> > > > $ ps aux | grep pgpool
> > > > ...
> > > > yugo-n   32156  0.0  0.0  19880  1244 pts/27   S    16:31   0:00
> pgpool:
> > > watchdog
> > > > yugo-n   32165  0.0  0.0      0     0 pts/27   Z    16:31   0:00
> > > [pgpool] <defunct>
> > > > yugo-n   32166  0.0  0.0  17792   768 pts/27   S    16:31   0:00
> pgpool:
> > > lifecheck
> > > > ...
> > > >
> > > > I found PID 32165 was pid of the escalation process.
> > > >
> > > >  2015-12-21 16:31:33: pid 32156: LOG:  escalation process started
> with
> > > PID:32165
> > > >  2015-12-21 16:31:33: pid 32165: LOG:  watchdog: escalation started
> > > >
> > > > I can reproduce this on Ubuntu but not CentOS.
> > > >
> > > > Regarads,
> > > > --
> > > > Yugo Nagata <nagata at sraoss.co.jp>
> > > > _______________________________________________
> > > > pgpool-hackers mailing list
> > > > pgpool-hackers at pgpool.net
> > > > http://www.pgpool.net/mailman/listinfo/pgpool-hackers
> > >
> > >
> > > --
> > > Yugo Nagata <nagata at sraoss.co.jp>
> > >
> > > _______________________________________________
> > > pgpool-hackers mailing list
> > > pgpool-hackers at pgpool.net
> > > http://www.pgpool.net/mailman/listinfo/pgpool-hackers
> > >
> > >
>
>
> --
> Yugo Nagata <nagata at sraoss.co.jp>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-hackers/attachments/20160105/5aed1719/attachment-0001.html>


More information about the pgpool-hackers mailing list