-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
4.18
-
Quality / Stability / Reliability
-
False
-
-
None
-
Important
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
-
None
Description of problem:
Clarify ETCD quorum restoration procedure
Version-Release number of selected component (if applicable):
At 4.18 OCP documentation, point 6.3.2. Quorum restoration stanza declares that:
When quorum is lost, the OpenShift Container Platform API becomes read-only.
But when quorum is lost the API is not available therefore not reachable at all, so the statement before seems not accurate.
Also at the 1st step of the procedure is pointing to perform verification to the API and it's not possibles as API is not available, so this need to be clarify.
How reproducible:
Check the quorum restoration as per documentation
Steps to Reproduce:
1. Shutdown 2 member of the 3 member that are part of the quorum
2. API is not available
3. connect by ssh on the only available member and execute the quorum_restore.sh
4. API is available again to start with the troubleshoot the etcd state.
Actual results:
Procedure at documentation is not accurate
Expected results:
Follow the documentation with the correct procedure to apply to recover the cluster
