[Pgpool-general] pgpool-ha and pg9 Streaming replication issue

[Yoshiharu Mori]

Takatuka san

> pgpool-ha has no special features for HS/SR, but it will work.
> 
> > > However pgpool does not like to promote the slave to primary in this
> > > case.
> 
> Does it mean that pgpool does not start on Node1 ?
> Then it is a problem heartbeat config or pgpool-ha. check heartbeat log.

Wouter uses pgpool-II 3.1.0-alpha1. 
Don't confuse by pgpool-II 3.0.1 and pgpool-II 3.1.0. 

> 2011-02-01 16:48:05 LOG:   pid 7837: pgpool-II successfully started.
> version 3.1.0-alpha1 (umiyameboshi)

pgpool-ha + pgpool-II 3.0.1+HS/SR is working except handling pgpool_status,
but I think pgpool-ha + pgpool-II 3.1 (with new fearures) is not work as he reported.  

It is necessary to discuss pgpool-HA and pgpool-II 3.1. 

On Thu, 3 Feb 2011 13:03:15 +0900
TAKATSUKA Haruka <harukat at sra.co.jp> wrote:

> 
> pgpool-ha has no special features for HS/SR, but it will work.
> 
> > > However pgpool does not like to promote the slave to primary in this
> > > case.
> 
> Does it mean that pgpool does not start on Node1 ?
> Then it is a problem heartbeat config or pgpool-ha. check heartbeat log.
> 
> ______________________________________________________________________
>  harukat at sraoss.co.jp  SRA OSS, Inc  http://www.sraoss.co.jp
> 
> 
> On Wed, 02 Feb 2011 08:35:35 +0900 (JST)
> Tatsuo Ishii <ishii at sraoss.co.jp> wrote:
> 
> > Maybe pgpool-ha is not yet ready for streaming replication.
> > Haruka?
> > --
> > Tatsuo Ishii
> > SRA OSS, Inc. Japan
> > English: http://www.sraoss.co.jp/index_en.php
> > Japanese: http://www.sraoss.co.jp
> > 
> > > Hi all,
> > > 
> > > I have 2 servers running postgres9 with streaming replication.
> > > On both nodes I have configured pgpool which are managed by pgpool-ha
> > > using heartbeat.
> > > 
> > > Now let's assume the following 
> > > 
> > > Node 0 is primary DB and running pgpool2
> > > Node 1 is standby DB
> > > 
> > > Now let's say I pull out the power for node0
> > > Heartbeat detects that node down and will fail-over the Virtual-IP and
> > > the pgpool2 instance as expected.
> > > 
> > > However pgpool does not like to promote the slave to primary in this
> > > case.
> > > 
> > > The log will show the following:
> > > 
> > > 2011-02-01 16:48:05 LOG:   pid 7837: pgpool-II successfully started.
> > > version 3.1.0-alpha1 (umiyameboshi)
> > > 2011-02-01 16:48:05 LOG:   pid 7837: find_primary_node: 1 node is
> > > standby
> > > 2011-02-01 16:48:05 LOG:   pid 7837: find_primary_node: no primary node
> > > found
> > > 
> > > Shouldn't pgpool2 promote the standby to primary as soon as he does not
> > > detect a primary node in the cluster.
> > > 
> > > I was thinking to run 2 instances at once and just used the Virtual-IP,
> > > however I'm not sure if this will work since these instances do not know
> > > about each other.
> > > 
> > > Thanks 
> > > 
> > > Wouter
> 
> _______________________________________________
> Pgpool-general mailing list
> Pgpool-general at pgfoundry.org
> http://pgfoundry.org/mailman/listinfo/pgpool-general
> 


-- 
Yoshiharu Mori <y-mori at sraoss.co.jp>