Skip to content

Commit ecdc871

Browse files
committed
Improvements based on @rguske feedback
1 parent 9c3e0dc commit ecdc871

8 files changed

Lines changed: 14 additions & 11 deletions

File tree

content/control-plane/node-failure/index.md renamed to content/control-plane/lost-quorum/index.md

Lines changed: 13 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -1,14 +1,17 @@
11
---
2-
title: Node failure test
3-
linktitle: Node failure
2+
title: Restoring etcd quorum (lost quorum)
3+
linktitle: Restoring etcd quorum (lost quorum)
44
description: Let's switch off some control plane nodes.
55
tags: ['etcd','control-plane','v4.17']
66
---
77

88
# Control plane node failure test
99

10-
* Just some stupid tests with OpenShift 4.17
11-
* **With [OCPSTRAT-539](https://issues.redhat.com/browse/OCPSTRAT-539) there will be a improvement of the process. Hopefully, land in 4.18!**
10+
Let's test how to restore a control-plane from a lost quorum. Losing the majority of control plane nodes leads to a quorum loss
11+
12+
Tested with OpenShift Cluster Version 4.17.0.
13+
14+
**With [OCPSTRAT-539](https://issues.redhat.com/browse/OCPSTRAT-539) there will be a improvement of the process. Hopefully, land in 4.18!**
1215

1316
Useful etcd commands
1417

@@ -62,7 +65,7 @@ Test Workload
6265
+---------------------------+------------------+---------+---------+-----------+------------+-----------+------------+--------------------+--------+
6366
```
6467

65-
## One control plane node stopeed
68+
## One control plane node stopped
6669

6770
Date CET: `2025-01-08 15:06:37 +0100`
6871

@@ -133,7 +136,7 @@ Test Workload
133136
stormshift-ocp1 #
134137
```
135138

136-
## Two control plane nodes stopeed
139+
## Two control plane nodes stopped
137140

138141
Date CET: `2025-01-08 15:16:06 +0100`
139142

@@ -156,7 +159,7 @@ Test Workload
156159
|---|---|---|---|---|
157160
|`2025-01-08 13:51:23 +0000`|🟢 🔴|🟢 🟢*|🟢 🟢|🟢 🟢|
158161

159-
* OpenShift Web console is "available" but without API useless and not really available.
162+
* OpenShift Web console is available. But the React application cannot connect to Kubernetes/OpenShift API any more.
160163
* Control plane is read-only == offline / not available
161164
* Workload is still running as expected
162165

@@ -308,7 +311,7 @@ Important points:
308311
[root@ocp1-cp-1 ~]#
309312
```
310313

311-
Accordint to monitoring, API was online at `2025-01-08 15:44:01`
314+
According to monitoring, API was online at `2025-01-08 15:44:01`
312315

313316
??? quote "oc get nodes (stormshift-ocp1)"
314317

@@ -413,7 +416,7 @@ Contiue with steps in docs (kubelet restart, ovn-kubernetes, csr,...)
413416
* ✅ DNS (A / PTR) done
414417
* ✅ DHCP done
415418
* ✅ RHCOS Live ISO Uploaded (rhcos-417.94.202410090854-0-live.x86_64.iso)
416-
* ✅ control plane igntion exported and available at `http://10.32.96.31/stormshift-ocp1-cp.ign`
419+
* ✅ control plane ignition file exported and available at `http://10.32.96.31/stormshift-ocp1-cp.ign`
417420

418421
### Add cp-4
419422

@@ -677,7 +680,7 @@ Patch object in another terminal
677680
stormshift-ocp1 #
678681
```
679682

680-
Let's delete the old control-plane artifacts
683+
Delete the old control-plane artifacts.
681684

682685
#### Delete two old control plane artifacts
683686

File renamed without changes.

content/control-plane/node-failure/ocp1-cp-4-machine.yaml renamed to content/control-plane/lost-quorum/ocp1-cp-4-machine.yaml

File renamed without changes.
File renamed without changes.
File renamed without changes.

content/control-plane/node-failure/ocp1-cp-5-machine.yaml renamed to content/control-plane/lost-quorum/ocp1-cp-5-machine.yaml

File renamed without changes.
File renamed without changes.

mkdocs.yml

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -174,7 +174,7 @@ nav:
174174

175175
- Control plane:
176176
- control-plane/index.md
177-
- Node failure: control-plane/node-failure/index.md
177+
- Restoring etcd quorum: control-plane/lost-quorum/index.md
178178

179179
- Troubleshooting: troubleshooting/index.md
180180

0 commit comments

Comments
 (0)