postgresql - No new leader elected after patroni node goes down

Question

I created a postgres HA cluster on Kubernetes with the pgo util from crunchydata. I used the following command:

pgo create cluster test --database newdb --sync-replication --memory=12Gi --pvc-size=20Gi --replica-count=1 --pgbouncer --pgbadger --metrics 

As expected, this created one leader and one replication node. To test the failover, I manually stopped and restarted the leader. After this, I have two replication nodes in the cluster. As there is no longer a leader, pgbouncer annot connect anymore.

Is this expected behaviour? I can manually start a leader election by calling

pgo failover mk --target test

But I would like this to happen automatically. Am I doing something wrong? Or is this expected to happen? I guess this should be same for all patroni clusters, independent from the crunchydata implementation

question from:https://stackoverflow.com/questions/65926226/no-new-leader-elected-after-patroni-node-goes-down

Answer 1

Waitting for answers