I have a master-slave Postgres setup in basic streaming replication I'm trying to get behind pgpool-II.
I have servers replicating find and connections to pgpool look to be getting load balanced between the two.
However in tests where I kill the primary in attempts to trigger failover_command it doesn't look to be electing the secondary. There are variables in the docs of "%m for new master id" and "%H for hostname of new master"
This are getting passed into the callback script with values "-1" and "0" respectively.
How do I get pgpool to recognize the slave as the "old master" so the failover command callback can finish appropriate actions for the promotion?