[pgpool-general: 2624] Re: pgpoll failure

[Yugo Nagata]

On Tue, 11 Mar 2014 12:02:26 -0300
Gonzalo Gil <gonxalo2000 at gmail.com> wrote:

> it's done now.
> 
> My problem now is that i am in a prod ambient... so i can play very much...
> 
> i have:
> node1, primary pgpool (with virtual ip) and primary database
> node2, standby node (database and pgpool)
> 
> i add a node (2nd standby in standby server) /datastb in pgpool.conf (bouth
> nodes)
> 
> from node1:
> # pcp_node_info 10 localhost 9898 postgres postgres 0
> prod 5432 *3* 1.000000
> # pcp_node_info 10 localhost 9898 postgres postgres 1
> replica 5432 *1* 0.000000
> # pcp_node_info 10 localhost 9898 postgres postgres 2
> replica 5444 3 0.000000
> 
> why node 0 (local) is donw!?

pgpool reads a file containing backend status at startup. pgpool at node #0
might read an old status file. To avoid this, start pgpool with -D option
and pgpool will discards the old status file.