Delete Percona Operator for PostgreSQL¶

When cleaning up your Kubernetes environment (e.g., moving from a trial deployment to a production one, or testing experimental configurations), you may need to remove some (or all) of the following objects:

Percona Distribution for PosgreSQL cluster managed by the Operator
Percona Operator for PostgreSQL itself
Custom Resource Definition deployed with the Operator

Delete a database cluster¶

You can delete the Percona Distribution for PosgreSQL cluster managed by the Operator by deleting the appropriate Custom Resource.

Note

There are two finalizers defined in the Custom Resource, which define whether TLS-related objects and data volumes should be deleted or preserved when the cluster is deleted.

finalizers.percona.com/delete-ssl: if present, objects, created for SSL (Secret, certificate, and issuer) are deleted when the cluster deletion occurs.
finalizers.percona.com/delete-pvc: if present, Persistent Volume Claims for the database cluster Pods are deleted when the cluster deletion occurs.

Both finalizers are off by default in the deploy/cr.yaml configuration file, and this allows you to recreate the cluster without losing data, credentials for the system users, etc.

Here’s a sequence of steps to follow:

List Custom Resources, replacing the <namespace> placeholder with your namespace.

$ kubectl get pg -n <namespace>

Sample output

NAME       ENDPOINT                         STATUS   POSTGRES   PGBOUNCER   AGE
cluster1   cluster1-pgbouncer.default.svc   ready    3          3           30m

Delete the Custom Resource with the name of your cluster (for example, let’s use the default cluster1 name).
```
$ kubectl delete pg cluster1 -n <namespace>
```
Sample output
```
perconapgcluster.pgv2.percona.com "cluster1" deleted
```
Check that the cluster is deleted by listing the available Custom Resources once again.
```
$ kubectl get pg -n <namespace>
```
Sample output
```
No resources found in <namespace> namespace.
```

Delete the Operator¶

You can uninstall the Operator by deleting the Deployments related to it.

List the deployments. Replace the <namespace> placeholder with your namespace.

$ kubectl get deploy -n <namespace>

Sample output

NAME                          READY   UP-TO-DATE   AVAILABLE   AGE
percona-postgresql-operator   1/1     1            1           13m

Delete the percona-* deployment

$ kubectl delete deploy percona-postgresql-operator -n <namespace>

Check that the Operator is deleted by listing the Pods. As a result you should have no Pods related to it.
```
$ kubectl get pods -n <namespace>
```
Sample output
```
No resources found in <namespace> namespace.
```

Delete Custom Resource Definition¶

If you are not just deleting the Operator and PostgreSQL cluster from a specific namespace, but want to clean up your entire Kubernetes environment, you can also delete the CustomRecourceDefinitions (CRDs) .

Warning

CRDs in Kubernetes are non-namespaced but are available to the whole environment. This means that you shouldn’t delete CRD if you still have the Operator and database cluster in some namespace.

You can delete CRD as follows:

List the CRDs:

$ kubectl get crd

Sample output

allowlistedv2workloads.auto.gke.io                   2023-09-07T14:15:30Z
allowlistedworkloads.auto.gke.io                     2023-09-07T14:15:29Z
audits.warden.gke.io                                 2023-09-07T14:15:32Z
backendconfigs.cloud.google.com                      2023-09-07T14:15:41Z
capacityrequests.internal.autoscaling.gke.io         2023-09-07T14:15:25Z
frontendconfigs.networking.gke.io                    2023-09-07T14:15:41Z
managedcertificates.networking.gke.io                2023-09-07T14:15:41Z
memberships.hub.gke.io                               2023-09-07T14:15:30Z
perconapgbackups.pgv2.percona.com                    2023-09-07T14:28:59Z
perconapgclusters.pgv2.percona.com                   2023-09-07T14:29:02Z
perconapgrestores.pgv2.percona.com                   2023-09-07T14:29:03Z
postgresclusters.postgres-operator.crunchydata.com   2023-09-07T14:29:06Z
serviceattachments.networking.gke.io                 2023-09-07T14:15:44Z
servicenetworkendpointgroups.networking.gke.io       2023-09-07T14:15:43Z
storagestates.migration.k8s.io                       2023-09-07T14:15:53Z
storageversionmigrations.migration.k8s.io            2023-09-07T14:15:53Z
updateinfos.nodemanagement.gke.io                    2023-09-07T14:15:55Z
volumesnapshotclasses.snapshot.storage.k8s.io        2023-09-07T14:15:52Z
volumesnapshotcontents.snapshot.storage.k8s.io       2023-09-07T14:15:52Z
volumesnapshots.snapshot.storage.k8s.io              2023-09-07T14:15:52Z

Now delete the percona*.pgv2.percona.com CRDs:

$ kubectl delete crd perconapgbackups.pgv2.percona.com perconapgclusters.pgv2.percona.com perconapgrestores.pgv2.percona.com

Sample output

customresourcedefinition.apiextensions.k8s.io "perconapgbackups.pgv2.percona.com" deleted
customresourcedefinition.apiextensions.k8s.io "perconapgclusters.pgv2.percona.com" deleted
customresourcedefinition.apiextensions.k8s.io "perconapgrestores.pgv2.percona.com" deleted

Get expert help¶

If you need assistance, visit the community forum for comprehensive and free database knowledge, or contact our Percona Database Experts for professional support and services. Join K8S Squad to benefit from early access to features and “ask me anything” sessions with the Experts.

Community Forum Get a Percona Expert Join K8S Squad

Last update: 2024-04-17