[pgpool-general: 2538] Re: Common practices for pgpool availability and data replication

[Gintautas Sulskus]

Thank you Tatsuo. That pretty much sums up the situation.

Gintas



On Tue, Feb 4, 2014 at 12:30 AM, Tatsuo Ishii <ishii at postgresql.org> wrote:

> I don't think it's possible because of PostgreSQL's limitation:
>
> In streaming replication, no standby server accepts DMLs in the first
> place.
>
> >>  *   synchronous replication:
> >>     *   If DML query relates to any of the table from that list, then
> >> pgpool would use synchronous replication - send DML to all DBs;
>
> So this part is not possible.
>
> Best regards,
> --
> Tatsuo Ishii
> SRA OSS, Inc. Japan
> English: http://www.sraoss.co.jp/index_en.php
> Japanese: http://www.sraoss.co.jp
>
> > We were considering the option to implement this functionality. Do you
> > think such configuration (my previous mail) is feasible on top of the
> > current pgool code base? Would be it difficult to implement?
> >
> > Cheers,
> > Gintas
> >
> >
> > On Mon, Feb 3, 2014 at 1:33 PM, Lazaro Ruben Garcia Martinez <
> > lgarciam at vnz.uci.cu> wrote:
> >
> >> An hybrid replication is not possible in this moment. Pgpool works in
> one
> >> specific mode, master-slave mode or syncronous replication.
> >>
> >> Regards.
> >> ________________________________________
> >> De: Gintautas Sulskus [gingaz at gmail.com]
> >> Enviado el: domingo, 02 de febrero de 2014 14:05
> >> Para: Lazaro Ruben Garcia Martinez
> >> CC: Athanasios | ZenGuard; pgpool-general at pgpool.net
> >> Asunto: Re: [pgpool-general: 2505] Re: Common practices for pgpool
> >> availability and data replication
> >>
> >> Thank you.
> >>
> >> Probably pgpool watchdog is the way to go for the first attempt.
> >>
> >> Regarding the second point. I had thoughts.. Would be possible to add a
> >> hybrid replication, that is:
> >>
> >>  *   pgpool would have a list of tables that must be synchronously
> >> replicated;
> >>  *   synchronous replication:
> >>     *   If DML query relates to any of the table from that list, then
> >> pgpool would use synchronous replication - send DML to all DBs;
> >>  *   asynchronous replication:
> >>     *   if DML query does not relate to any table from the list, then it
> >> would be sent to the master database and be replicated via streaming
> >> replication
> >>
> >> Is such approach technically feasible (if I wanted to add it)? Would it
> be
> >> difficult?
> >> I just had a thought, what if pgpool replicates DML query to all
> servers.
> >> How master's streaming replication then behaves?
> >>
> >> Cheers,
> >> Gintas
> >>
> >>
> >> On Thu, Jan 30, 2014 at 2:18 PM, Lazaro Ruben Garcia Martinez <
> >> lgarciam at vnz.uci.cu<mailto:lgarciam at vnz.uci.cu>> wrote:
> >> For number one you should use watchdog functionality.
> >>
> >> Regards.
> >> ________________________________________
> >> De: pgpool-general-bounces at pgpool.net<mailto:
> >> pgpool-general-bounces at pgpool.net> [pgpool-general-bounces at pgpool.net
> >> <mailto:pgpool-general-bounces at pgpool.net>] En nombre de Athanasios |
> >> ZenGuard [athanasios at zenguard.org<mailto:athanasios at zenguard.org>]
> >> Enviado el: jueves, 30 de enero de 2014 6:48
> >> Para: Gintautas Sulskus
> >> CC: pgpool-general at pgpool.net<mailto:pgpool-general at pgpool.net>
> >> Asunto: [pgpool-general: 2505] Re: Common practices for pgpool
> >> availability and data replication
> >>
> >> Hi,
> >> for point 1) you can use keepalived and two pgpool machines
> >>
> >>
> >> On Thu, Jan 30, 2014 at 12:00 PM, Gintautas Sulskus <gingaz at gmail.com
> >> <mailto:gingaz at gmail.com><mailto:gingaz at gmail.com<mailto:
> gingaz at gmail.com>>>
> >> wrote:
> >> Dear All,
> >>
> >> I have a few questions regarding practices with pgpool:
> >> 1. Typically pgpool is a single point of failure. What is the common
> >> practice to solve this?
> >> 2. Some of our data (e.g. 10% tables) must be consistent across the
> >> cluster for every transaction. The rest may differ. Is it possible to
> have
> >> a fine-grained replication, where some DMLs would be replicated
> >> synchronously and the rest would use asynchronous replication?
> >>
> >> Thanks!
> >>
> >> Best Regards,
> >> Gintautas
> >>
> >> _______________________________________________
> >> pgpool-general mailing list
> >> pgpool-general at pgpool.net<mailto:pgpool-general at pgpool.net><mailto:
> >> pgpool-general at pgpool.net<mailto:pgpool-general at pgpool.net>>
> >> http://www.pgpool.net/mailman/listinfo/pgpool-general
> >>
> >>
> >>
> >> ________________________________
> >> Fin a la injusticia, LIBERTAD AHORA A NUESTROS CINCO COMPATRIOTAS QUE SE
> >> ENCUENTRAN INJUSTAMENTE EN PRISIONES DE LOS EEUU!
> >> http://www.antiterroristas.cu
> >> http://justiciaparaloscinco.wordpress.com
> >>
> >> Fin a la injusticia, LIBERTAD AHORA A NUESTROS CINCO COMPATRIOTAS QUE SE
> >> ENCUENTRAN INJUSTAMENTE EN PRISIONES DE LOS EEUU!
> >> http://www.antiterroristas.cu
> >> http://justiciaparaloscinco.wordpress.com
> >>
> >>
> >>
> >> --
> >> Best Regards,
> >> Gintautas Sulskus
> >>
> >> ________________________________
> >> Fin a la injusticia, LIBERTAD AHORA A NUESTROS CINCO COMPATRIOTAS QUE SE
> >> ENCUENTRAN INJUSTAMENTE EN PRISIONES DE LOS EEUU!
> >> http://www.antiterroristas.cu
> >> http://justiciaparaloscinco.wordpress.com
> >>
> >> Fin a la injusticia, LIBERTAD AHORA A NUESTROS CINCO COMPATRIOTAS QUE SE
> >> ENCUENTRAN INJUSTAMENTE EN PRISIONES DE LOS EEUU!
> >> http://www.antiterroristas.cu
> >> http://justiciaparaloscinco.wordpress.com
> >>
> >
> >
> >
> > --
> > Best Regards,
> > Gintautas Sulskus
>



-- 
Best Regards,
Gintautas Sulskus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20140206/97dd4b6b/attachment.html>