If i understand your scenario correct, after failover, Patroni created deleted files on old primary by replciating from New primary?
If that is correct, i would recommend to check lag between new primary and old primary(now slave). if it is zero then we are good to perform failover.
Regards,
Abdul Sayeed
On Wed, Jul 20, 2022 at 5:29 PM Sankar, Uma (Uma) <usankar@avaya.com> wrote:
Hi All,
This is regarding the Postgres HA working with patroni in 3 node setup, we have an issue with the primary because a few database files were deleted manually so performed a switch over to move the services from primary to secondary with patroni, post the switchover was deleted file was re-created by itself in the old primary.
Now we are planning to switch back the services from secondary(Leader) to primary(replica) as the deleted files were re-created itself, can someone please suggest if this works normally when we switch back to promote primary as the leader and secondary as a replica.
