[pgpool-general: 1916] Re: database recovery with a pgpool-II cluster

Tatsuo Ishii ishii at postgresql.org
Tue Jul 23 08:17:29 JST 2013


> Hi,
> 
> http://www.pgpool.net/docs/latest/pgpool-en.html says this:
> 
> "While executing online recovery in replication mode, you need to stop
> all pgpool-II instances except the one which is doing online
> recovery. Otherwise DB can be running into inconsistent state."
> 
> I have a pair of clustered pgpool-II instances (and two PostgreSQL
> servers) using the watchdog and the latest code from git.  Does the
> above advice still apply?  I seem to recall someone saying the
> restriction would be relaxed in 3.3.

Yes, with 3.2 or 3.3 you don't need to worry about it.

> Also, with clustered pgpool-II instances it appears to be impossible
> to keep identical configuration files, because wd_hostname needs to be
> the external name of the host (as opposed to, say, "localhost"). Is
> that right?  Or is there some way for the configurations to be made
> identical?

You are correct. the watchdog section in pgpool.conf cannot be
identical.

I will commit a doc update.
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese: http://www.sraoss.co.jp


More information about the pgpool-general mailing list