Major upgrades single-instance vs. cluster

[uk1988]

postgres-operator version v1.13.0
Clarification on the major upgrade process needed

Case 1 - DB Cluster
We have a PostgreSQL cluster with two instances running PostgreSQL version 14. We update the PostgreSQL manifest to set the version to 15. Within a few minutes, the postgres-operator detects the manifest change and updates the cluster node by node. After the process completes, I SSH into the primary pod and run SELECT version(); in psql, which confirms the PostgreSQL version is 15.

Case 2 - Single Pod
We repeat the same process on a single-instance PostgreSQL deployment. The operator detects the manifest change, and the single pod is restarted. However, running SELECT version(); still returns PostgreSQL version 14. If I manually terminate the pod after the restart, the version remains 14. However, if I delete the pod, an internal Python script runs, and the version updates to 15.

Questions:

• Why does the upgrade behavior differ between the cluster and single-node setups?
• How can I force the upgrade on a single-node setup?

@CyberDem0n i have seen your PR from 2020.
There is no indication that there is diff between 1 and 2 pg pods upgrade.

[uk1988]

After detail look at operator log I can see that when we have a single node setup upgrade of the DB is stopped.
The only way is to create a script that will run the upgrade command that is in the container.