Postgres Operator Multi-Cluster auto failover

[iosifidd]

Description:

Hello,
we are evaluating using the Percona PostgreSQL operator for deploying PostgreSQL in an active-standby topology, needed for HighAvailability. I have seen the guide for achieving this:

Percona Operator for PostgreSQL - Introduction

but as i see, it is not possible to perform an automatic failover. Meaning, it is documented how to promote a standby cluster to active, but it is not clear what to do to the previous active cluster. There is also a reference on the documentation:

Percona Operator for PostgreSQL - Introduction

but it is quite vague.

Any advice?

Thank you!

[Agustin_G]

Hi @iosifidd,

The operator uses Patroni Standby cluster functionality, so its limitations apply here too. You can read more about it in the Patroni documentation:

HA multi datacenter — Patroni 4.0.5 documentation

Once you do the standby cluster promotion, you need to make sure that no traffic is going to the previous primary cluster.

Regarding:

Meaning, it is documented how to promote a standby cluster to active, but it is not clear what to do to the previous active cluster.

This will vary depending on each environment, and the steps needed are covered to some extent in the Patroni link above.
This is also explained in the Percona operator documentation link:

Percona Operator for PostgreSQL - Failover

Lastly, you sent the same link twice, so I’m not sure what your second reference was pointing to.