I have a Postgres pod which has a mounted volume:

kind: PersistentVolume
apiVersion: v1
metadata:
  name: postgres-pv
  labels:
    type: local
spec:
  storageClassName: manual
  capacity:
    storage: 100M
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/mnt/data"
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  labels:
    app: postgres
  name: psql-claim
spec:
  storageClassName: manual
  accessModes:
  - ReadWriteOnce
  resources:
    requests:
      storage: 100M

The yaml contains the following:

volumeMounts:
            - mountPath: /docker-entrypoint-initdb.d/
              name: psql-config-map-volume
            - mountPath: /var/lib/postgresql/data
              name: psql-claim 
              subPath: postgres
      volumes:
        - name: psql-config-map-volume
          configMap:
            name: psql-config-map // contains an init.sql
        - name: psql-claim
          persistentVolumeClaim:
            claimName: psql-claim

It works well, data retain after deployment/pod deletion and redeploy.

The problem appears when I modify the init.sql. It didn't come into effect and got this message at psql pod startup:

PostgreSQL Database directory appears to contain a database; Skipping initialization

The pod itself starts without any error (just with the old init.sql data)

What I have tried:

1. Deleting the deployment,the pvc and pv. Then redeploy everything. Same result.

2. I searched the hostpath data but /mnt/data is empty.

What else should I try? How can I force the init.sql to launch? Where the actual old data is stored if not in the hostpath?

edit: I have search for pg files and found this:
/var/lib/docker/overlay2/6ae2../merged/mnt/data/postgres/pg_ident.conf
/var/lib/docker/overlay2/6ae2../diff/mnt/data/postgres/pg_ident.conf
And it still exists after pvc and pv deletion. How can I gracefully reset its data?