Backing Up and Restoring VMware SQL with MySQL for Kubernetes Instances

This topic describes how to back up and restore VMware SQL with MySQL for Kubernetes.

Overview

VMware MySQL Operator allows you to generate on-demand backups, configure schedules for automated backups, and restore backups to new MySQL instances.

For uploading and retrieving backup artifacts, VMware MySQL Operator currently supports S3-compatible storage and Azure Blob Storage.

For backing up and restoring, VMware MySQL Operator uses four of the five Custom Resource Definitions (CRDs):

MySQLBackup: References a MySQL backup artifact that exists in an external blobstore such as S3 or Minio. A new MySQLBackup resource is created every time an on-demand or scheduled backup is generated.
MySQLBackupLocation: References an external blobstore and credentials necessary to access the blobstore.
MySQLBackupSchedule: Represents a CronJob schedule on which to perform backups.
MySQLRestore: References an instance of a restore that was performed. A new MySQLRestore resource is created every time a restore is performed.

For detailed information about the CRDs, see Controllers and Custom Resource Definitions (CRDs) in Architecture.

Note: The procedures in this topic require the VMware SQL with MySQL for Kubernetes deployment templates. If you have not yet downloaded the deployment templates, see Download the Deployment Templates in Creating and Deleting MySQL Instances.

About Synchronization of Backups with the External Blobstore

VMware MySQL Operator syncs MySQLBackup resources in a Kubernetes cluster with the contents of the external blobstore. The external blobstore is treated as the source of truth. This means that, if a MySQLBackup resource is deleted on the Kubernetes cluster, but the associated backup artifact still exists in the external blobstore, VMware MySQL Operator re-creates the MySQLBackup resource to match the contents of the external blobstore.

Back Up VMware MySQL Operator Data

Performing backups for VMware MySQL Operator requires creating a MySQLBackupLocation resource that references an external blobstore. Both on-demand backups and scheduled backups use the MySQLBackupLocation to upload backup artifacts to the external blobstore.

Before starting the procedures for backing up a VMware MySQL Operator instance, ensure that you know the configuration details of your external blobstore and how often you want to perform scheduled backups.

Create a MySQLBackupLocation Resource

The MySQLBackupLocation Resource is used to configure the namespace with the location of the blobstore, and the credentials to access it. If your storage endpoint uses custom TLS certificates, provide your CA bundle in the MySQLBackupLocation Resource.

For details on the backup location CR and its properties, refer to Properties for the MySQLBackupLocation Resource, and Properties for the Secret.

To create a MySQLBackupLocation resource:

Find the backuplocation.yaml deployment template that you downloaded in the TGZ file from VMware Tanzu Network. For how to download deployment templates, see Download the Deployment Templates in Creating and Deleting MySQL Instances.

Create a copy of the backuplocation.yaml file and give it a unique name.

For example:

cp ~/Downloads/mysql-for-kubernetes-1.7.1/samples/backuplocation.yaml testbackuplocation.yaml

Edit the file with the configuration details of your external blobstore.
Create the MySQLBackupLocation resource in the same namespace as the MySQL instances that you want to back up by running:
```
kubectl apply -f FILENAME -n DEVELOPMENT-NAMESPACE
```
- Where is DEVELOPMENT-NAMESPACE is the namespace for the MySQL instance.
- Where FILENAME is the name of the configuration file you created in Step 2 above.
For example:
```
kubectl apply -f testbackuplocation.yaml -n my-namespace
```
```
mysqlbackuplocation.with.sql.tanzu.vmware.com/backuplocation-sample created
secret/backuplocation-sample-creds configured
```

Verify that the MySQLBackupLocation has been created by running:

kubectl get mysqlbackuplocation backuplocation-sample \
-o jsonpath={.spec} -n DEVELOPMENT-NAMESPACE

For example:

kubectl get mysqlbackuplocation backuplocation-sample -o jsonpath={.spec} -n my-namespace

{
  "storage": {
    "s3": {
      "bucket": "bucket-sample",
      "forcePathStyle": false,
      "region": "us-west-1",
      "secret": {
        "name": "backuplocation-sample-creds"
      }
    }
  }
}

Create a MySQLBackupSchedule Resource

To set a schedule for automatic backups, create a MySQLBackupSchedule resource:

Find the backupschedule.yaml deployment template that you downloaded in the TGZ file from VMware Tanzu Network. For how to download deployment templates, see Download the Deployment Templates in Creating and Deleting MySQL Instances.

Create a copy of the backupschedule.yaml file and give it a unique name.

For example:

cp ~/Downloads/mysql-for-kubernetes-1.7.1/samples/backupschedule.yaml testbackupschedule.yaml

Edit the file with the name of the MySQLBackupLocation resource that you created in Create a MySQLBackupLocation Resource and the name of the MySQL instance you want scheduled backups of. For an explanation of the properties that you can set in this file, see Properties for the MySQLBackupSchedule Resource.
Create the MySQLBackupSchedule resource in the same namespace as the MySQLBackupLocation and MySQL instance that you referenced in the MySQLBackupSchedule YAML file.
```
kubectl apply -f FILENAME -n DEVELOPMENT-NAMESPACE
```
Where FILENAME is the name of the configuration file you created in Step 2 above.

For example:
```
kubectl apply -f testbackupschedule.yaml -n my-namespace
```
```
mysqlbackupschedule.with.sql.tanzu.vmware.com/backupschedule-sample created
```

Verify that the MySQLBackupSchedule has been created by running:

kubectl get mysqlbackupschedule mysqlbackupschedule-sample -o jsonpath={.spec} -n DEVELOPMENT-NAMESPACE

For example:

kubectl get mysqlbackupschedule mysqlbackupschedule-sample -o jsonpath={.spec} -n my-namespace
{
  "backupTemplate": {
    "spec": {
      "instance": {
        "name": "demo-db"
      },
      "location": {
        "name": "demo-backuplocation"
      }
    }
  },
  "schedule": "@daily"
}

If you correctly configured both a MySQLBackupLocation resource and MySQLBackupSchedule resource for an existing MySQL instance, you see backups being generated and uploaded to the external blobstore.

Name and Location for Backup Artifacts

MySQLBackup resources that are automatically generated as a result of a MySQLBackupSchedule are named SCHEDULE-NAME-TIMESTAMP.

By default, VMware MySQL Operator stores backup artifacts under the subfolder structure yyyy > mm > dd. You can configure a custom path for backups so that backup artifacts are stored under the subfolder structure CUSTOM-PATH > yyyy > mm > dd.

Backup artifacts stored in the external blobstore are named DATETIME-RANDOM_STRING-backup.xb.

For example, if a MySQLBackupSchedule name is mysqlbackupschedule-sample, the custom backup path is my-backups/, and a backup was taken on Thursday, December 10, 2020 at 8:51:03 PM GMT (timestamp 1607633463), then:

The MySQLBackup resource on the Kubernetes cluster is named mysqlbackupschedule-sample-1607633463
The backup artifact in the external blobstore is named 20201210T205103-kzw54l-backup.xb
The path to the artifact is my-backups/2020/12/10/.

Perform an On-Demand Backup

In addition to scheduled backups, you can take individual backups whenever you want.

Prerequisite: A MySQLBackupLocation resource that represents the external blobstore to which you upload the generated backup artifact. To configure the MySQLBackupLocation resource, see Create a MySQLBackupLocation Resource above.

To take a backup:

Find the backup.yaml deployment template that you downloaded in the TGZ file from VMware Tanzu Network. For how to download deployment templates, see Download the Deployment Templates in Creating and Deleting MySQL Instances.

Create a copy of the backup.yaml file and give it a unique name.

For example:

cp ~/Downloads/mysql-for-kubernetes-1.7.1/samples/backup.yaml testbackup.yaml

Edit the file. For an explanation of the properties that you can set for the MySQLBackup resource, see Properties for the MySQLBackup Resource.
Trigger the backup by creating the MySQLBackup resource in the same namespace as the instance by running:
```
kubectl apply -f FILENAME -n DEVELOPMENT-NAMESPACE
```
Where FILENAME is the name of the configuration file you created in Step 2 above.

For example:
```
kubectl apply -f testbackup.yaml -n my-namespace
mysqlbackup.with.sql.tanzu.vmware.com/backup-sample created
```

Verify that a backup has been generated and track its progress by running:

kubectl get mysqlbackup backup-sample -n DEVELOPMENT-NAMESPACE

For example:

kubectl get mysqlbackup backup-sample -n my-namespace

NAME            STATUS      SOURCE INSTANCE   TIME STARTED           TIME COMPLETED
backup-sample   Succeeded   mysql-sample      2020-12-01T21:49:26Z   2020-12-01T21:49:30Z

For an explanation of what each column means, see List Existing MySQLBackup Resources below.

List Existing MySQLBackup Resources

You might want to list existing MySQLBackup resources for various reasons, for example:

To select a backup to restore. For steps to restore a backup, see Restore.
To see the last successful backup.
To verify that scheduled backups are running as expected.
To find old backups that need to be cleaned up. For steps to delete backups, see Delete Old Backup Artifacts.

To see a list of existing MySQLBackup resources:

List existing MySQLBackup resources by running:

kubectl get mysqlbackup

For example:

kubectl get mysqlbackup
NAME            STATUS   SOURCE INSTANCE   TIME STARTED   TIME COMPLETED
backup-sample   Failed   mysql-sample

To understand the output, see the table below:

Column Name	Meaning
`STATUS`	Represents the current status of the backup. Allowed values are: Pending: The backup has been received but not scheduled on a MySQL Pod. Running: The backup is being generated and streamed to the external blobstore. Succeeded: The backup has completed successfully. Failed: The backup has failed to complete. To troubleshoot a failed backup, see Troubleshoot Backup and Restore below.
`SOURCE INSTANCE`	The MySQL instance the backup was taken from.
`TIME STARTED`	The time that the backup process started.
`TIME COMPLETED`	The time that the backup process finished. If the backup fails, this value is empty.

Delete Old Backup Artifacts

VMware SQL with MySQL for Kubernetes does not natively support retention policies for backup artifacts. You can configure retention policies on your external blobstore. If you do, you must also delete the associated MySQLBackup resources in the Kubernetes cluster, because those are not automatically deleted by VMware SQL with MySQL for Kubernetes.

To delete a backup:

Delete the backup in the external blobstore.

On your Kubernetes cluster, delete the MySQLBackup resource by running:

kubectl delete mysqlbackup BACKUP-NAME -n DEVELOPMENT-NAMESPACE

For example:

kubectl delete mysqlbackup backup-sample -n my-namespace

Restore

This section discusses two kinds of restore:

Restore from a Backup
Restore a Backup to a Different Namespace or Kubernetes Cluster

Restore from a Backup

MySQLRestores always restores to a new MySQL instance to avoid overwriting any data on an existing MySQL instance. When the restore is triggered, it automatically creates the new MySQL instance.

VMware MySQL Operator does not allow you to restore a backup to an existing MySQL instance. Although you can perform this manually by copying the MySQL data from the backup artifact onto an existing MySQL instance, VMware strongly discourages you from doing this because you might overwrite existing data on the MySQL instance.

Prerequisites

Before you restore from a backup, you must have:

An existing MySQLBackup in your current namespace. To select an existing MySQLBackup to restore, see List Existing MySQLBackup Resources.
A MySQLBackupLocation that represents the bucket where the existing backup artifact is stored. See Create a MySQLBackupLocation Resource above.

Procedure

To restore from a backup:

Find the restore.yaml deployment template that you downloaded in the TGZ file from VMware Tanzu Network. For how to download deployment templates, see Download the Deployment Templates in Creating and Deleting MySQL Instances.

Create a copy of the restore.yaml file and give it a unique name.

For example:

cp ~/Downloads/mysql-for-kubernetes-1.7.1/samples/restore.yaml testrestore.yaml

Edit the file. For information about the properties that you can set for the MySQLRestore resource, see Property Reference for Backup and Restore.

To restore from an HA instance to a non-HA instance, edit the testrestore.yaml and amend the spec.instanceTemplate.spec.highAvailability field to false. For example:

---
apiVersion: with.sql.tanzu.vmware.com/v1
kind: MySQLRestore
metadata:
  name: restore-sample
spec:
  backup:
     name: backup-sample
  instanceTemplate:
     metadata:
       name: mysql-sample
     spec:
       storageSize: 1Gi
       imagePullSecretName: tanzu-image-registry
       highAvailability:
         enabled: false

Trigger the restore by creating the MySQLRestore resource in the same namespace as the MySQLBackup and MySQLBackupLocation by running:
```
kubectl apply -f FILENAME -n DEVELOPMENT-NAMESPACE
```
Where FILENAME is the name of the configuration file you created in Step 2 above.

For example:
```
kubectl apply -f testrestore.yaml -n my-namespace
```
```
mysqlrestore.with.sql.tanzu.vmware.com/restore-sample created
```

Verify that a restore has been triggered and track the progress of your restore by running:

kubectl get mysqlrestore restore-sample -n DEVELOPMENT-NAMESPACE

For example:

kubectl get mysqlrestore restore-sample -n my-namespace

NAME             STATUS      SOURCE BACKUP   TARGET INSTANCE   TIME STARTED           TIME COMPLETED
restore-sample   Succeeded   backup-sample   mysql-sample      2020-12-01T21:52:30Z   2020-12-01T21:53:09Z

To understand the output, see the table below:

Column Name	Meaning
`STATUS`	Represents the current status of the restore process. Allowed values are: Pending: The restore has been received but not yet scheduled on a MySQL Pod. Running: The restore is in progress. Succeeded: The restore has completed successfully. Failed: The restore failed. To troubleshoot, see Troubleshoot Backup and Restore below.
`SOURCE BACKUP`	The name of the backup being restored.
`TARGET INSTANCE`	The name of the new MySQL instance to be restored with the backup contents.
`TIME STARTED`	The time that the restore process started.
`TIME COMPLETED`	The time that the restore process finished. If the restore fails, this value is empty.

Restoring a Backup to a Different Namespace or Kubernetes Cluster

If you want to restore a backup to a different namespace or a different Kubernetes cluster, create a MySQLBackupLocation in the target namespace or Kubernetes cluster. Then, VMware MySQL Operator automatically creates MySQLBackup resources for the backup artifacts in the external blobstore.

To restore to a different namespace or Kubernetes cluster, you create a BackupLocation in the target namespace:

Target the destination Kubernetes cluster or namespace.
Create a MySQLBackupLocation resource that contains the backup artifact to restore. For how to do this, see Create a MySQLBackupLocation Resource.

Confirm that the MySQLBackup artifact to restore is included in the list by running:

kubectl get mysqlbackup

For example:

kubectl get mysqlbackup

NAME            STATUS      SOURCE INSTANCE   TIME STARTED           TIME COMPLETED
sample-backup   Succeeded   mysql-sample      2020-12-01T21:49:26Z   2020-12-01T21:49:30Z

Trigger a restore by following steps in Restore from a Backup.

Troubleshoot Backup and Restore

Basic troubleshooting begins with reviewing the status for the resource and reading the messages associated with the resource events.

To troubleshoot problems with backup and restore:

Detect issues by monitoring the STATUS column of any MySQL custom resource. If the status is Failed or is stuck in Pending, Scheduled, or Running, then one of the following might be the problem:
- Misconfiguration
- Problem with the external blobstore
- Issues with the MySQL Operator
In this example, the kubectl get command outputs a Failed status:
```
kubectl get mysqlbackup backup-sample
```
```
NAME            STATUS   SOURCE INSTANCE   TIME STARTED   TIME COMPLETED
backup-sample   Failed   mysql-sample
```

Diagnose the issue by inspecting the Kubernetes events for the resource, for example:

kubectl get events --field-selector involvedObject.name=backup-sample

LAST SEEN   TYPE      REASON   OBJECT                      MESSAGE
2m43s       Warning   Failed   mysqlbackup/backup-sample   Secret "backuplocation-sample-creds" not found

Read the message in the MESSAGE column to understand why the failure occurred.

In the example above, the backup-sample expected a Kubernetes secret called backuplocation-sample-creds to exist. Fix this problem by creating the backuplocation-sample-creds secret. The template for this secret is located in the backuplocation.yaml template.