Here’s a realistic Azure-specific DR test using
OpenShift Container Platform +
OpenShift API for Data Protection (OADP).

We’ll simulate a namespace + data loss and walk through a full restore using Azure Blob + Disk snapshots.

Scenario (Azure DR test)

❌ my-app namespace deleted
❌ PVC + data gone
❌ Need full recovery from backup

Environment:

• OADP configured with Azure Blob
• CSI snapshots enabled (Azure Disk)

What we’re restoring

• Kubernetes resources (deployments, services, routes)
• Persistent volumes (via Azure snapshots)
• Application data

Flow overview

Backup (Blob + Disk Snapshot)
        ↓
Namespace deleted ❌
        ↓
Velero Restore triggered
        ↓
Resources recreated
        ↓
PVC restored from snapshot
        ↓
App back online ✅

Step-by-step restore

Step 1: Confirm failure

oc get ns my-app

Should show:

NotFound

Step 2: List available backups

oc get backup -n openshift-adp

Example:

azure-backup   Completed

Step 3: Create restore

apiVersion: velero.io/v1
kind: Restore
metadata:
  name: restore-my-app
  namespace: openshift-adp
spec:
  backupName: azure-backup
  includedNamespaces:
    - my-app

Apply:

oc apply -f restore.yaml

Step 4: Watch restore progress

oc get restore -n openshift-adp

Detailed:

oc describe restore restore-my-app -n openshift-adp

Step 5: Verify namespace restored

oc get ns my-app

Then:

oc get pods -n my-app

Step 6: Verify PVC restoration

oc get pvc -n my-app

Check:

• Status = Bound

Step 7: Verify Azure disk restore

In Azure:

az disk list --resource-group <rg>

You should see:

• restored disk from snapshot

Step 8: Check application

oc get routes -n my-app

Test:

curl http://<route>

What just happened

1. OADP pulled metadata from Azure Blob
2. Recreated Kubernetes objects
3. Triggered Azure disk snapshot restore
4. Reattached volumes to pods

Full app recovery

Real-world variations

Case 1: Partial restore

Restore only one resource:

includedResources:
  - deployments

Case 2: Restore to different namespace

namespaceMapping:
  my-app: my-app-restore

Case 3: Restore without volumes

restorePVs: false

Azure-specific pitfalls

1. Missing snapshot permissions

→ restore fails silently or PVC stuck

2. Storage class mismatch

→ PVC stays Pending

3. Region mismatch

→ snapshot cannot attach

4. Private cluster networking

→ cannot reach Blob storage

Troubleshooting

Check restore logs

oc logs -n openshift-adp deployment/velero

Check events

oc get events -n my-app

Check PVC issues

oc describe pvc <pvc-name> -n my-app

Pro DR test (recommended)

Simulate:

1. Backup app
2. Delete namespace
3. Restore
4. Validate data integrity

Do this quarterly

Advanced Azure DR test

Try:

• Restore to new cluster in different region
• Reconnect DNS
• Validate external integrations

Key takeaway

• Azure DR = Blob (metadata) + Disk snapshot (data)
• OADP restores both together
• Works for full or partial recovery