[pgpool-general: 2923] Re: pgpool does not restart child processes after failback (pcp_attach_node)

[Guy Meler]

Oh, I see...:)
Thank you for your comment Tatsuo.
Wouldn't it be better that this behavior will be configurable via the conf
file?

Thank you.


On Wed, Jun 18, 2014 at 12:43 PM, Tatsuo Ishii <ishii at postgresql.org> wrote:

> > Hey :)
> > I have pgpool with 2 backends, master/slave streaming replication mode.
> > When I stop the service on the slave, the node is deattched correctly and
> > all child processes are restarted to update thier backends connections [
> > when i run show pool_pools I can see that every pool_pid was changed and
> > all pools are connected to one backend, the master ].
> > However,
> > when I start the service again on the failed standby, and run the
> > pcp_attach_node command, the child processes are not restarted, thus, all
> > connections that were open before the failback, are remain open and
> > connected only to the master, until a pgpool restart is invoked and then,
> > all the pools connections are connected to both backends. I need to
> mention
> > that every *new* session to the pgpool after the attache, is connected to
> > both backends.
> >
> > My question is, why after pcp_attach_node command, the child processes
> are
> > not restarted as they should, like they do after failover.
> >
> > Thank you very much!
>
> Actually pgpool-II behaved as you wish before. However lots of users
> complained that already connected sessions were disconnected when
> standby nodes were re-attached. Thus I fixed things as users
> requested.
>
> Best regards,
> --
> Tatsuo Ishii
> SRA OSS, Inc. Japan
> English: http://www.sraoss.co.jp/index_en.php
> Japanese:http://www.sraoss.co.jp
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20140618/2ef6f46b/attachment.html>