Skip to main content
RunBook Academy

← All runbooks in Ceph

medium risksecurity relevant~45 min

Rotate a cephx key

1 · Prerequisites

Confirm every item is in place before any state change.

  • A rotation cadence is scheduled (quarterly is recommended)
  • The cluster is HEALTH_OK
  • The keyring is current

2 · Pre-checks

Read-only diagnostic commands. If any of these don't match expected output, stop and investigate further.

  • · ceph auth list shows the entity to rotate
  • · The new keyring distribution target is identified
  • · The application has been notified of the rotation
  • · A maintenance window is scheduled (if the application cannot take a key change live)

3 · Procedure

Execute each step in order. Verify the expected output of a step before moving to the next.

  1. 1Generate the new key: ceph auth get-or-create-key client.<user>
  2. 2Distribute the new keyring to the legitimate consumers
  3. 3Wait for the application to use the new key (or restart it)
  4. 4Confirm both keys are accepted: ceph auth export client.<user> > /tmp/old-keyring.bin
  5. 5Set the old key to expire (or remove): ceph auth rm client.<user> --old
  6. 6Verify the application continues to work after the old key is removed

4 · Verification

Confirm the procedure actually fixed the problem.

  • The new key is in use by the application
  • The old key is removed
  • The application continues to work
  • The rotation is documented

5 · Rollback

If verification fails, undo the procedure in reverse order.

  • If the application cannot use the new key: distribute the prior key back as a fallback
  • If the cluster breaks: investigate which consumer was using the old key
  • If the rotation cannot be verified: redo with a clear schedule

6 · Escalation

When the runbook isn't enough, contact:

  • · Security team
  • · Application team
  • · Storage team

Rotate a cephx key

This runbook covers Rotate a cephx key on a Ceph Tentacle cluster.

When to use

  • Active production scenario matching the impact (security-relevant)

Prerequisites

  • A rotation cadence is scheduled (quarterly is recommended)
  • The cluster is HEALTH_OK
  • The keyring is current

Pre-checks

# ceph auth list shows the entity to rotate
# The new keyring distribution target is identified
# The application has been notified of the rotation
# A maintenance window is scheduled (if the application cannot take a key change live)

Procedure

Step 1

Generate the new key: ceph auth get-or-create-key client.<user>

Step 2

Distribute the new keyring to the legitimate consumers

Step 3

Wait for the application to use the new key (or restart it)

Step 4

Confirm both keys are accepted: ceph auth export client.<user> > /tmp/old-keyring.bin

Step 5

Set the old key to expire (or remove): ceph auth rm client.<user> —old

Step 6

Verify the application continues to work after the old key is removed

Verification

# The new key is in use by the application
# The old key is removed
# The application continues to work
# The rotation is documented

Rollback

  • If the application cannot use the new key: distribute the prior key back as a fallback
  • If the cluster breaks: investigate which consumer was using the old key
  • If the rotation cannot be verified: redo with a clear schedule

Escalation

  • Security team
  • Application team
  • Storage team

References

  1. Ceph documentation