Velero Reports Partial Failure "Error backing up item error: pod volume backup failed: error to initialize data path: error to boost backup repository connection"

Products

• VMware Tanzu Kubernetes Grid Management

Issue/Introduction

Velero backup job reports status PartiallyFailed due to consistent failures in backing up persistent volume data via PodVolumeBackups.

• Metadata objects (Deployments, Services, Secrets, PVC references) are successfully backed up, reflected in the high item count (e.g., 3951).

  velero backup describe <backup-name> --details

  Status:
  Completion Timestamp:  2025-07-04T23:32:12Z
  Errors:                33
  Expiration:            2025-08-03T19:31:49Z
  Format Version:        1.1.0
  Phase:                 PartiallyFailed
  Progress:
    Items Backed Up:     3951
    Total Items:         3951

• Pod volume data is excluded from the final backup due to repository initialisation issues. The below error can be seen in the describe of velero backup
  
  error to initialize data path: error to boost backup repository connection default-tanzu-system-registry-kopia: error to connect backup repo: error to connect repo with storage: error to connect to repository: repository not initialized in the provided storage"

• Error logs show repeated Kopia repo failures across multiple namespaces

  kubectl logs -n velero -l name=velero --tail=1000 | grep 'pod volume backup failed'

  Error backing up item: pod volume backup failed: 
error to initialize data path: error to boost backup repository connection 
"<kopia-repo-name>": error to connect backup repo: 
error to connect repo with storage: error to connect to repository: 
repository not initialized in the provided storage

Environment

VMware Tanzu Kubernetes Grid Management
Velero 1.15+

Cause

Velero backup jobs report PartiallyFailed status due to repeated PodVolumeBackup failures tied to Kopia repository initialization issues. Although metadata objects (Deployments, Services, PVC references, etc.) are backed up successfully, the backup job excludes pod volume data because the target repository is either uninitialized or unreachable.

Resolution

To prevent failures like `repository not initialized in the provided storage`, perform the following checks and initialization steps before triggering backups.

• If using Kopia (recommended from Velero 1.15 onward), please ensure the repository is initialised:

• Validate Repository Status

  velero repo get --backup-location default  # Replace 'default' with your backup-location name

status: Ready
repoIdentifier: default-<namespace>-kopia

  If the repository status is empty or not ready try reinitializing
   
  velero repository create --backup-location <your-location>
• If Restic is still in use, verify the repository exists at the configured storage path.
  Note: Restic is deprecated in Velero 1.15+.
• Check that the Velero node-agent DaemonSet is healthy and scheduled on all nodes:
• kubectl get pods -n velero -o wide kubectl describe daemonset velero -n velero
• Once the above steps are performed, retry the backup and see if the velero backup status is success. If not reach out to Broadcom support team

Additional Information