Recover a sealed cluster · 0.5.x

A running but sealed Pod cannot serve requests. Identify the configured seal first, then repair its existing credential, trust, or network path. Do not create new seal material as a trial fix.

Confirm the failure

inspect

Read seal and dependency signals

kubectl -n <namespace> get openbaocluster <name> \
  -o jsonpath='{range .status.conditions[*]}{.type}={.status} {.reason}{"\n"}{.message}{"\n"}{end}'
kubectl -n <namespace> get openbaocluster <name> -o yaml | yq '.spec.unseal'
kubectl -n <namespace> exec <pod-name> -- bao status
kubectl -n <namespace> logs <pod-name> -c openbao

OpenBaoSealed is derived from OpenBao’s service-registration label on Pods; use bao status to confirm. If the cluster is unsealed but has no leader, switch to Recover from no leader.

Diagnose the selected seal

Seal	Check first
`static` or omitted	Exact generated `<cluster>-unseal-key` Secret, its ownership, and its `key` data
`transit`	Credentials Secret, transit policy and key, HTTPS trust and SNI, DNS, and egress
`awskms`, `gcpckms`, `azurekeyvault`, `ocikms`	`CloudUnsealIdentityReady`, workload ServiceAccount binding, decrypt permission, DNS, and egress
`kmip`	Client certificate, key and CA files, server name, and endpoint reachability
`pkcs11`	Vendor library and dependent library paths, token or slot, PIN material, and device access

For external seals, compare the Secret keys and mounted file paths with Configure unseal. Look for permission denied, x509, AccessDenied, and timeout errors before rotating any credential.

Recover the static-seal Secret

The default static seal uses an operator-generated immutable Secret named <cluster>-unseal-key. Its key contains 32 random bytes mounted at the configured static-seal path.

Inspect metadata and presence without printing Secret data into terminal history:

inspect

Check the static-seal Secret

kubectl -n <namespace> get secret <name>-unseal-key -o json | \
  jq '{name: .metadata.name, immutable: .immutable, keys: (.data | keys)}'

If the Secret is missing, pause the cluster and restore the original Secret with the ownership metadata required by the target OpenBaoCluster. If that material is lost, a Raft snapshot alone does not recreate the missing seal key.

Verify recovery

After repairing the dependency, restart through spec.runtime.restartAt only if the process must reload mounted material. Then verify every voter:

verify

Confirm unseal and service state

kubectl -n <namespace> get pods -l openbao.org/cluster=<name>
kubectl -n <namespace> exec <pod-name> -- bao status
kubectl -n <namespace> exec <pod-name> -- bao operator raft list-peers

raft list-peers requires an authenticated OpenBao session with Raft configuration read access. Use your approved interactive login path; do not put a privileged token in the command line or shell history.

Finish when all intended voters are unsealed and Ready, a leader exists, and Available=True. A manual bao operator unseal prompt is not a generic escape hatch for the operator’s auto-seal modes; repair the configured seal or recover its original material.