Hello,
Did anybody solve a problem of logical replication slot fail-over on a streaming replication cluster?
I have a logical decoding on a logical replication slot set up on a streaming replication master.
Streaming replication does not propagate the replication slot state. In case of fail-over, I need to create a new replication slot on a new muster and loose some database changes.
Is there any work-around that would allow avoid loosing database changes on a logical replication slot after a fail-over?
Postgres 10.8 on CentOS 7.
--
Thanks
Igor Polishchuk