Uploaded image for project: 'OpenShift API for Data Protection'
  1. OpenShift API for Data Protection
  2. OADP-3759

Velero shouldn't restore the restore-wait init container

XMLWordPrintable

    • 4
    • False
    • Hide

      None

      Show
      None
    • False
    • ToDo
    • 0
    • 0.000
    • Very Likely
    • 0
    • None
    • Unset
    • Unknown
    • No

      Description of problem:

      I executed the native datamover backup and restore on same namespace after successful backup and restore via Kopia.  I noticed velero is restoring restore-wait init container even though it's not required in this case. 

      Version-Release number of selected component (if applicable):
      OADP 1.3.1-49

       

      How reproducible:
      Always

       

      Steps to Reproduce:
      1. Deploy an application consisting PVC's

      2. Trigger a file system backup
      3. Delete app namepace

      4. Execute Restore 
      5. Now backup the same namespace via native datamover or CSI
      6. Delete app namespace
      7. Execute restore

       

      Actual results:

      Application pod has the init container added in it. 

      $ oc get pod -n ocp-mysql
       initContainers:
          - args:
            - 926e0ef5-5038-49d4-8f63-259acee86e42
            command:
            - /velero-restore-helper
            env:
            - name: POD_NAMESPACE
              valueFrom:
                fieldRef:
                  apiVersion: v1
                  fieldPath: metadata.namespace
            - name: POD_NAME
              valueFrom:
                fieldRef:
                  apiVersion: v1
                  fieldPath: metadata.name
            image: registry.redhat.io/oadp/oadp-velero-restic-restore-helper-rhel9@sha256:d2c7da625f2d1c6d54f6eb62d247acdc542e3a065f7e89c62e371b946a77c7e5
            imagePullPolicy: IfNotPresent
            name: restore-wait

      Expected results:

      Velero should skip the restoring of initContainer spec. 

       

      Additional info:

      I encountered an issue were my application pod was stuck at init container phase. I don't think it's an usual scenario just adding here for the awareness.

      $ oc get pod -n ocp-mysql
      NAME                     READY   STATUS     RESTARTS   AGE
      mysql-7f77d47fd7-694pn   0/1     Init:0/1   0          35m
      
      $ oc get pvc -n ocp-mysql
      NAME      STATUS   VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS   AGE
      mysql     Bound    pvc-279cf584-f50a-4859-83fd-3354150278ec   2Gi        RWO            standard-csi   36m
      mysql-1   Bound    pvc-df0ebad2-cb36-4e62-9eff-6e4f06be980f   2Gi        RWO            standard-csi   36m
      
      $ oc describe pod -n ocp-mysql
      Events:
        Type    Reason                  Age   From                     Message
        ----    ------                  ----  ----                     -------
        Normal  Scheduled               35m   default-scheduler        Successfully assigned ocp-mysql/mysql-7f77d47fd7-694pn to oadp-74411-q4bpw-worker-c-z6m65
        Normal  SuccessfulAttachVolume  35m   attachdetach-controller  AttachVolume.Attach succeeded for volume "pvc-df0ebad2-cb36-4e62-9eff-6e4f06be980f"
        Normal  SuccessfulAttachVolume  35m   attachdetach-controller  AttachVolume.Attach succeeded for volume "pvc-279cf584-f50a-4859-83fd-3354150278ec"
        Normal  AddedInterface          35m   multus                   Add eth0 [10.131.0.23/23] from ovn-kubernetes
        Normal  Pulled                  35m   kubelet                  Container image "registry.redhat.io/oadp/oadp-velero-restic-restore-helper-rhel9@sha256:d2c7da625f2d1c6d54f6eb62d247acdc542e3a065f7e89c62e371b946a77c7e5" already present on machine
        Normal  Created                 35m   kubelet                  Created container restore-wait
        Normal  Started                 35m   kubelet                  Started container restore-wait 

       
      Steps I followed. 
      1. Deployed an application with PVCs

      2. First I triggered the FileSystem Backup with kopia
      3. Removed the app namespace and executed the restore
      4. Again I backed up the same namespace with CSI

      5. Followed the same procedure delete the app namespace and ran restore.
      6. Again I backed the same namespace with native dataMover.
      7. Deleted the app namespace and ran restore. 

       

      Conversation on forum-oadp https://redhat-internal.slack.com/archives/C0144ECKUJ0/p1710940866491269

              wnstb Wes Hayutin
              rhn-support-prajoshi Prasad Joshi
              Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

                Created:
                Updated: