[pgpool-general: 5752] Re: pg_terminate_backend -> postmaster on DB node 1 was shutdown by administrative command

Alberto Rodríguez García albertorodriguezgarcia at gmail.com
Tue Sep 26 05:38:31 JST 2017


Hello,

After more investigation and Log reading we found out that pgpool dropping
the master was caused by a fast shutdown request on the master, done while
failed backup attempt.

Not pgpool issue after all.

Kind regards.

Alberto Rodriguez.
Amsterdam, Nederland.

On Mon, Sep 25, 2017 at 6:15 PM, Alberto Rodríguez García <
albertorodriguezgarcia at gmail.com> wrote:

> Hello,
>
> I think that my issue is related to Avi's issue.
>
> I am running a multi data center setup with pgpool-II version 3.6.2
> (subaruboshi) and postgresql 9.6.2
>
> 2 datacenter + 1 datacenter as witness.
>
> 2 pgsql per datacenter and 2 pgpool per data center sharing 1 VIP per
> datacenter.
>
> I am using pgpool,the streaming replication from postgresql and repmgr for
> master failover and witness server.
>
> Last friday I got this from different pgpool services in different
> datacenters but using same master backend:
>
> ....
> 2017-09-22 18:26:40: pid 15131: LOG:  received degenerate backend request
> for node_id: 2 from pid [15131]
> 2017-09-22 18:26:40: pid 7215: LOG:  watchdog received the failover
> command from local pgpool-II on IPC interface
> 2017-09-22 18:26:40: pid 7215: LOG:  watchdog is processing the failover
> command [DEGENERATE_BACKEND_REQUEST] received from local pgpool-II on IPC
> interface
> 2017-09-22 18:26:40: pid 7215: LOG:  forwarding the failover request
> [DEGENERATE_BACKEND_REQUEST] to all alive nodes
> 2017-09-22 18:26:40: pid 7215: DETAIL:  watchdog cluster currently has 1
> connected remote nodes
> 2017-09-22 18:26:40: pid 7215: LOG:  new IPC connection received
> 2017-09-22 18:26:40: pid 7215: LOG:  watchdog received the failover
> command from local pgpool-II on IPC interface
> 2017-09-22 18:26:40: pid 7215: LOG:  watchdog is processing the failover
> command [DEGENERATE_BACKEND_REQUEST] received from local pgpool-II on IPC
> interface
> 2017-09-22 18:26:40: pid 7215: LOG:  failover command
> [DEGENERATE_BACKEND_REQUEST] from local pgpool-II on IPC interface is
> ignored
> ....
>
> The result of this was that pgpool removed the master from the load
> balancer, leaving only the standby, so the setup was in "read-only".
>
> I have attached part of the log file for more information and also I
> include the pgpool.conf
>
> What can I do to prevent this situation with pgpool?
>
> Thank you
>
> Kind regards.
>
> Alberto Rodriguez.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20170925/cac48a7c/attachment.html>


More information about the pgpool-general mailing list