I have installed and created a default PostgreSQL cluster with kind “PerconaPGCluster”. My question is:
-Is it already equipped with High availability feature? Or should I add additional config (e.g. Patroni) to achieve HA feature?
Thank you in advance!
Hi @Luan ,
After you deploy the
PerconaPGCluster you don’t need to do anything for HA, Patroni is already configured and deployed by the operator. You can read more in the docs. Also you might be interested in this webinar.
Thank you very much for your quick nice answer.
In my test, I created a cluster with default configuration. After few seconds, I see there are 1 primary pods and 2 replicas created and associated with the new cluster.
Then I connected my web application with it, and everything works well (I created some application users, then login the app with them).
Then I manually deleted the running primary Postgres pod instance. I see the database is not accessible for about 1 minute (until I see a new primary pod instance created).
Is the “1 minute” expected downtime? is that due to failing over process or is that the time for re-creating a new primary pod instance?
Thank you again!
Yes, and yes. K8S cannot “instantly” recreate a new pod with all provisioned storage etc.