After deploying a Ceph cluster, a systematic validation pass confirms the cluster is correctly configured before it handles production workloads. Catching misconfiguration early is far less disruptive than finding it under load.

Check 1 - Cluster Health

ceph status
ceph health detail

Expected result: HEALTH_OK. Investigate any warnings before proceeding.

Check 2 - MON Quorum

ceph quorum_status --format json-pretty | python3 -c "
import sys, json
q = json.load(sys.stdin)
print('Leader:', q['quorum_leader_name'])
print('Members:', q['quorum_names'])
"

Verify all expected MONs are in quorum.

Check 3 - OSD Count and Distribution

ceph osd tree

Verify:

• Correct number of OSDs are up and in
• OSDs are distributed across the expected hosts and failure domains
• Weights reflect the disk sizes

ceph osd df tree

Check 4 - CRUSH Map Topology

ceph osd crush tree

Confirm the CRUSH topology matches your intended failure domains (host, rack, datacenter).

Check 5 - Pool Replication Settings

ceph osd pool ls detail | grep -E "size|min_size|pg_num"

For each pool, verify:

• size = intended replication factor (e.g., 3)
• min_size = at least 2 for production
• pg_num appropriate for pool size and OSD count

Check 6 - Network Configuration

ceph config get mon public_network
ceph config get osd cluster_network
ceph quorum_status --format json-pretty | python3 -c "
import sys, json
q = json.load(sys.stdin)
for mon in q['monmap']['mons']:
    print(mon['name'], 'addr=', mon['addr'], 'public_addr=', mon['public_addr'])
"

Verify bindings are on the expected networks and not on 0.0.0.0 unless intended.

Check 7 - StorageClass Provisioning Test

kubectl apply -f - <<'EOF'
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: validation-pvc
spec:
  accessModes: [ReadWriteOnce]
  resources:
    requests:
      storage: 1Gi
  storageClassName: rook-ceph-block
EOF

kubectl wait --for=jsonpath='{.status.phase}'=Bound pvc/validation-pvc --timeout=60s
kubectl delete pvc validation-pvc

Check 8 - Read/Write Performance Baseline

TOOLBOX=$(kubectl get pod -n rook-ceph -l app=rook-ceph-tools -o name | head -1)
POOL=$(kubectl get storageclass rook-ceph-block -o jsonpath='{.parameters.pool}')
kubectl exec -n rook-ceph "$TOOLBOX" -- rados bench -p "$POOL" 30 write --no-cleanup
kubectl exec -n rook-ceph "$TOOLBOX" -- rados bench -p "$POOL" 30 seq
kubectl exec -n rook-ceph "$TOOLBOX" -- rados cleanup -p "$POOL"

Record this baseline for future comparison.

Summary

Validating a Ceph cluster after deployment requires checking cluster health, MON quorum, OSD topology, CRUSH map, pool settings, network bindings, PVC provisioning, and establishing a performance baseline. Running all eight checks takes about 15 minutes and catches the most common deployment errors before they affect production workloads.