Details
-
Improvement
-
Resolution: Unresolved
-
Major
-
None
-
None
-
1
Description
PDBs are great for ensuring that voluntary disruptions don't cause outages for Couchbase Server.
In pre-prod environments, these can be a bit of a pain in situations where the uptime of the database isn't really important.
They make the worker node upgrade process very slow, when the user can tolerate some downtime and just have the Operator spin up new pods on other worker nodes.
It would be good if as part of the CouchbaseCluster, you could opt out (so they're enabled by default) of using PDBs to simplify non-production deployments of Couchbase Server.
You could also check whether PVCs are enabled for the pods too, to ensure that there won't be unexpected data loss.