Ian thank you very much for the help.
I found the problem.
When an automated failover take places, the repmgr process on the failed
node dies.
Postgresql service stops on Node1 , Node2 is promoted as master
automaticaly, but repmgr on node1 stops, and I need to start repmgr again
after recover Node1.
Is this the expected behavior?
Regards.
-----Mensaje original-----
De: Ian Barwick [mailto:ian.barwick@2ndquadrant.com]
Enviado el: martes, 14 de febrero de 2017 02:55 a. m.
Para: Lazaro Garcia; pgsql-admin@postgresql.org
Asunto: Re: [ADMIN] repmgr error with automated failover
Hi
On 02/13/2017 11:32 PM, Lazaro Garcia wrote:
> Hello every one in the list. First thanks to repmgr.
>
>
>
> I took several days testing repmgr but I can't execute an automated
> failover properly.
>
>
>
> I have 2 nodes, a master(node1) and a slave(nodo2). The repmgrd
> service is running on both servers. When node1 goes down an automated
> failover take place and node2 is promounted as master, at this point
> all is ok, the problems comes when node2 is attached again to the
> cluster, for incorporating this node to the cluster I run this commands:
>
> 1. executes a clone command on node1
>
> su -c 'repmgr -h node2 -U repmgr -d repmgr -D
> /var/lib/postgresql/9.6/main/ -f /etc/postgresql/9.6/main/repmgr.conf
--rsync-only --force standby clone'
> postgres
>
>
>
> 2. start up the postgres service on node1
>
>
>
> 3. Register this node
>
> su -c 'repmgr -f /etc/postgresql/9.6/main/repmgr.conf --force standby
> register' postgres
>
>
>
> Then stops node2 but node1 is never promoted as master.
>
> Why node1 is not promoted? Is there something wrong?
Is repmgrd running on node1? Is its repmgr.conf file configured correctly,
e.g. "failover=automatic"
Regards
Ian Barwick
--
Ian Barwick http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services