[Pgpool-general] diagnosing BackendError from pcp_recovery_node

[Josh Berkus]

On 10/5/11 3:15 PM, Jeff Frost wrote:
> I have a follow_master_command set which calls the pcp_recovery_node script
> after editing a config file on the replica which has been degenerated. 
> Occasionally this fails with BackendError.  It looks like the node doesn't get
> degenerated fast enough and so it's marked as still alive, thus cancelling the
> recovery that I just requested:

Ishii-san, any thoughts on this?

-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com