[pgpool-general: 5018] Re: PgPool fails to restart while being master

[Jacobo García López de Araujo]

I have been able to find a workaround this issue by adding a function on
the init script that waits for the port 9009 to be close before starting
PgPool on the restart command of the script.







On Wed, Sep 21, 2016 at 7:49 PM Jacobo García López de Araujo <
jacobo.garcia at gmail.com> wrote:

> I am finding this trouble in a two nodes cluster setup. Both nodes have
> the same configuration (I can post it if needed). I can restart the standby
> node as many times I want without problem.
>
> When I try to restart the master node, it stops correctly but then when it
> starts again it brings the following error message (redacted)
>
> Sep 21 17:38:45 srv0.net pgpool: 2016-09-21 17:38:45: pid 1947: FATAL:
>  failed to create watchdog receive socket
> Sep 21 17:38:45 srv0.net pgpool: 2016-09-21 17:38:45: pid 1947: DETAIL:
>  bind on "TCP:9009" failed with reason: "Address already in use"
> Sep 21 17:38:45 srv0.net pgpool: 2016-09-21 17:38:45: pid 1940: LOG:
>  watchdog child process with pid: 1947 exits with status 768
> Sep 21 17:38:45 srv0.net pgpool: 2016-09-21 17:38:45: pid 1940: FATAL:
>  watchdog child process exit with fatal error. exiting pgpool-II
>
> Linked is the complete error message:
> https://gist.github.com/therobot/dc410f0ccc56e302866329c43670ec38
>
> It seems to me that PgPool fails to close the watchdog socket (9009) on
> time and when it started it finds the socket still opened.
>
> The init script sends a -INT signal to PgPool in order to stop it. I have
> also tried replacing the killproc signal with 'pgpool stop -m smart' in the
> init script without too much luck.
>
> Thanks in advance.
>
> --
> Jacobo García López de Araujo.
>
-- 
Jacobo García López de Araujo.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20160922/04cf06d3/attachment.html>