Backup and restore as code

oc create -f <storagelocation.yaml>

Sample storage location YAML file:

apiVersion: v1
data:
  access-key-id: AMIATNJ3JEMKR6GAUOLN
  secret-access-key: vgm9AJPztPkOygBFpBp2UzEErLBelTcp3JPdPn9c
kind: Secret
metadata:
  name: backup-storage-location-example-secret-0
  namespace: ibm-spectrum-fusion-ns
---
apiVersion: data-protection.isf.ibm.com/v1alpha1
kind: BackupStorageLocation
metadata:
  
  name: backup-storage-location-example
  namespace: ibm-spectrum-fusion-ns
spec:
  type: aws
  credentialName:  backup-storage-location-example-secret-0
  provider: isf-backup-restore
  params:
    region: us-west-2
    bucket: bucket-name
    endpoint: https://s3.us-west-2.amazonaws.com

Run the following command to generate the cloud data:

printf "[default]\naws_access_key_id=minio\naws_secret_access_key=minio123\n" | base64 -w 0

oc create -f <backuppolicy.yaml>

oc apply -f <backuppolicy.yaml>

Sample backup policy YAML files:

• Backup policy to take backups daily:

  
  apiVersion: data-protection.isf.ibm.com/v1alpha1
  kind: BackupPolicy
  metadata:
    name: daily-policy
    namespace: ibm-spectrum-fusion-ns
  spec:
    provider: isf-backup-restore
    backupStorageLocation: backup-storage-location-example
    retention:
      number: 10
      unit: days
    schedule:
      cron: "30 10 * * *"
      timezone: America/Los_Angeles
      

• Backup policy to take backups weekly from Monday through Friday:

  apiVersion: data-protection.isf.ibm.com/v1alpha1
  kind: BackupPolicy
  metadata:
    name: backup-policy-weekdays
    namespace: ibm-spectrum-fusion-ns
  spec:
    provider: isf-backup-restore
    backupStorageLocation: backup-storage-location-example
    retention:
      number: 4
      unit: weeks
    schedule:
      cron: "30 20 * * 1,2,3,4,5"
      timezone: America/Los_Angeles

oc delete backuppolicy <backuppolicy name>

Example:

oc delete backuppolicy daily-policy

oc describe backuppolicy daily-policy 

Alternatively, you can use the following sample OC command:

oc describe backuppolicy/daily-policy

oc create -f <policyassignment.yaml>

Sample assignment YAML file for daily backup:

apiVersion: data-protection.isf.ibm.com/v1alpha1
kind: PolicyAssignment
metadata:
  name: backup-policy-assignment-example
  namespace: ibm-spectrum-fusion-ns
spec:
  application: application-sample
  backupPolicy: daily-policy
  appCluster: fusion-cluster
  

Run the following command to get the name:

oc get cluster

Example:

apiVersion: data-protection.isf.ibm.com/v1alpha1
kind: PolicyAssignment
metadata:
  
  name: filebrowser-20220919-1834111-isf-auto-ibmcos-backuppolicy-20220919-183411
  namespace: ibm-spectrum-fusion-ns
spec:
  application: filebrowser-20220919-1834111
  backupPolicy: isf-auto-ibmcos-backuppolicy-20220919-183411
  appCluster: fusion-cluster
  runNow: true

The corresponding backup CRD file is as follows:

apiVersion: data-protection.isf.ibm.com/v1alpha1
kind: BackupPolicy
metadata:
  
  name: isf-auto-ibmcos-backuppolicy-20220919-183411
  namespace: ibm-spectrum-fusion-ns
spec:
  backupStorageLocation: isf-ibm-location-20220919-183411
  provider: isf-backup-restore
  retention:
    number: 1
    unit: days
  schedule:
    cron: 51 18 1 * *
    timezone: America/Los_Angeles

oc create -f <backup-restore.yaml>

For cross cluster restore, the targetCluster is provided. If targetCluster is not available, then default to the same cluster where the backup got originated from:

apiVersion: data-protection.isf.ibm.com/v1alpha1
kind: Restore
metadata:
  dp.isf.ibm.com/provider-name: isf-backup-restore
  name: restore-complete-wordpress-1689030046
  namespace: ibm-spectrum-fusion-ns
spec:
  backup: complete-wordpress-ibm-do-not-use-202307102253
  objectsToRestore:
    RESOURCES:
    - ALL
    v1/persistentvolumeclaim:
    - complete-wordpress/mysql-pv-claim
    - complete-wordpress/wp-pv-claim
  targetCluster: guard-vpc-dog-98b7318c91b01bd72490e80cc2328915-0000.us-south.containers.appdomain.cloud

For example, application A comprising of three namespaces namely ns1, ns2, and ns3 are backed up, then you need to specify the namespaceMapping field in Restore CR to restore this application to alternate namespace.

apiVersion: data-protection.isf.ibm.com/v1alpha1
kind: Restore
metadata:
  name: <custom-restore-name>
  namespace: ibm-spectrum-fusion-ns
spec:
  backup: <your-backup-job-name>
  namespaceMapping:
      ns1: ns1-new
      ns2: ns2-new
      ns3: ns3-new

Namespace variables automatically point to the correct namespaces during backup and restore. And it would be useful in scenarios, where you want to refer current namespace like exec hooks, mainly in restore cases.

An example of application that is backed up with three variables in Application CR.

apiVersion: application.isf.ibm.com/v1alpha1
kind: Application
...
spec:
  includedNamespaces:
    - ns1
    - ns2
    - ns3
  variables:
    - name: PARENT_NAMESPACE
      value: ns1
    - name: CHILD1_NAMESPACE
      value: ns2
    - name: CHILD2_NAMESPACE
      value: ns3

There are two methods to access these variables:

• The first method is using ${PARENT_NAMESPACE}. It is valid only in backup scenario, and becomes invalid during restore as still referring to ns1.
• The second method is using new construct ${APP.${PARENT_NAMESPACE}.name}, which substitutes namespace variables correctly. This refers to ns1 in case of backup and ns1-new when restoring to alternate namespace.

  In a restore scenario within the same namespace, it refers to ns1 only. This new construct can be utilized while specifying hook namespace field and exec commands.