public/rancher-docs
1
0
Fork 0
mirror of https://github.com/rancher/rancher-docs.git synced 2026-05-14 00:53:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add leader election troubleshooting docs

Browse Source
This commit is contained in:
Sebastiaan van Steenis
2018-12-27 11:14:25 +01:00
committed by Denise
parent 527fdc14c5
commit c161f88f83
3 changed files with 23 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
content/rancher/v2.x/en/troubleshooting/kubernetes-components/_index.md
+2
View File
@@ -254,6 +254,8 @@ bdf3898b8063 rancher/hyperkube:v1.11.5-rancher1 "/opt/rke-tools/en..."
### controlplane container logging
> **Note:** If you added multiple nodes with the `controlplane` role, both `kube-controller-manager` and `kube-scheduler` use a leader election process to determine the leader. Only the current leader will log the performed actions. See [Kubernetes leader election]({{< baseurl >}}/rancher/v2.x/en/troubleshooting/kubernetes-resources/#kubernetes-leader-election) how to retrieve the current leader.
The logging of the containers can contain information on what the problem could be.
```
content/rancher/v2.x/en/troubleshooting/kubernetes-resources/_index.md
+20
View File
@@ -50,6 +50,26 @@ Example output:
worker-0: DiskPressure:True
```
### Kubernetes leader election
#### Kubernetes Controller Manager leader
The leader is determined by a leader election process. After the leader has been determined, the leader (`holderIdentity`) is saved in the `kube-controller-manager` endpoint (in this example, `controlplane-0`).
```
kubectl -n kube-system get endpoints kube-controller-manager -o jsonpath='{.metadata.annotations.control-plane\.alpha\.kubernetes\.io/leader}'
{"holderIdentity":"controlplane-0_xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx","leaseDurationSeconds":15,"acquireTime":"2018-12-27T08:59:45Z","renewTime":"2018-12-27T09:44:57Z","leaderTransitions":0}>
```
#### Kubernetes Scheduler leader
The leader is determined by a leader election process. After the leader has been determined, the leader (`holderIdentity`) is saved in the `kube-scheduler` endpoint (in this example, `controlplane-0`).
```
kubectl -n kube-system get endpoints kube-scheduler -o jsonpath='{.metadata.annotations.control-plane\.alpha\.kubernetes\.io/leader}'
{"holderIdentity":"controlplane-0_xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx","leaseDurationSeconds":15,"acquireTime":"2018-12-27T08:59:45Z","renewTime":"2018-12-27T09:44:57Z","leaderTransitions":0}>
```
### Ingress Controller
The default Ingress Controller is NGINX and is deployed as a DaemonSet in the `ingress-nginx` namespace. The pods are only scheduled to nodes with the `worker` role.
content/rancher/v2.x/en/troubleshooting/networking/_index.md
+1 -1
View File
@@ -99,7 +99,7 @@ If there is no output, the cluster is not affected.
#### System namespace pods network connectivity broken
> Note: This applies only to Rancher upgrades from v2.0.6 or earlier to v2.0.7 or later. Upgrades from v2.0.7 to later version are unaffected.
> **Note:** This applies only to Rancher upgrades from v2.0.6 or earlier to v2.0.7 or later. Upgrades from v2.0.7 to later version are unaffected.
| | |
|------------|------------|