[Pgpool-general] Issue with master node selection in pgpool-II-3.0.3-beta1

[Alexander Zhuravlev]

Hello,

On Thu, Feb 24, 2011 at 03:43, Tatsuo Ishii <ishii at sraoss.co.jp> wrote:
> Zhuravlev,
>
> I found a bug possibly related to your report.  Pgpool worker process
> have a persistent connection to backends to check replication
> delay. If a backend goes down and then up, the process tries to keep
> using the connection without knowing the fact. I think this is the
> cause of the problems you have seen.
>
> The fix was included in 3.0.3, released yesterday (though no mention
> about this on the NEWS file).
>
> http://lists.pgfoundry.org/pipermail/pgpool-committers/2011-February/001669.html
>
> Now the process creates everytime when replication delay checked. This
> is add a little bit resource consumption, but I couldn't think of any
> alternative.
>
> Could you please test 3.0.3 and report back to me if your problems go
> away?

It looks like the fix has resolved the issue. I do not see the error
messages anymore.

Thank you for your help.
-- 
Alexander Zhuravlev
http://ikato.com/