Making on-demand backup¶

Before you begin¶

Export the namespace as an environment variable. Replace the <namespace> placeholder with your value:
```
export NAMESPACE = <namespace>
```
Check the configuration of the PerconaServerMySQL object:
- Check that the backup.enabled key is set to true. Use the following command:
```
kubectl get ps <cluster-name> -n $NAMESPACE -o jsonpath='{.spec.backup.enabled}'
```
- Verify that you have configured backup storage and specified its configuration in the backup.storages subsection of the Custom Resource.
For incremental backups, check implementation specifics and rules.

Backup steps¶

To make an on-demand backup, use a special backup configuration YAML file. The example of such file is deploy/backup/backup.yaml .

You can check available options in the Backup resource reference

Specify the following keys:

Set the metadata.name key to assign a name to the backup.
Set the spec.clusterName key to the name of your cluster.
Set the spec.storageName key to a storage configuration defined in your deploy/cr.yaml file.
Set the spec.type key to make an incremental backup. When unset, a full backup is made by default.
Optionally, specify the base backup for the spec.incrementalBaseBackupName. This option is only valid for the spec.type=incremental.
Optionally, add the percona.com/delete-backup entry under metadata.finalizers to enable deletion of backup files from a cloud storage when the backup object is removed (manually or by schedule).

Pass this information to the Operator:

via the YAML manifestvia the command line

Edit the deploy/backup/backup.yaml file:

apiVersion: ps.percona.com/v1
kind: PerconaServerMySQLBackup
metadata:
  name: backup1
  finalizers:
    - percona.com/delete-backup
spec:
  clusterName: ps-cluster1
  storageName: s3-us-west

Start the backup process:

kubectl apply -f deploy/backup/backup.yaml -n $NAMESPACE

Instead of storing backup settings in a separate file, you can pass them directly to the kubectl apply command as follows:

cat <<EOF | kubectl apply -n $NAMESPACE -f-
apiVersion: ps.percona.com/v1
kind: PerconaServerMySQLBackup
metadata:
  name: backup1
  finalizers:
    - percona.com/delete-backup
spec:
  clusterName: ps-cluster1
  storageName: s3-us-west
EOF

List backups with this command:

kubectl get ps-backup -n $NAMESPACE

Specifying the backup source¶

When you create a backup object, the Operator selects a Pod to take the backup from. You can see the backup source pod in the backup object’s status:

kubectl get ps-backup backup1 -o yaml

Sample output

apiVersion: ps.percona.com/v1
kind: PerconaServerMySQLBackup
metadata:
  name: backup1
  ...
status:
  backupSource: cluster1-mysql-1.cluster1-mysql.<namespace>
  ...

You can specify the source pod in the backup object to run the backup on this specific pod:

apiVersion: ps.percona.com/v1
kind: PerconaServerMySQLBackup
metadata:
  name: backup1
  finalizers:
    - percona.com/delete-backup
spec:
  clusterName: ps-cluster1
  storageName: s3-us-west
  sourcePod: ps-cluster1-mysql-2

Troubleshooting¶

If you face issues with backups, refer to our Backup troubleshooting guide for help.

Last update: April 17, 2026
Created: April 17, 2026