ansible-deployer:k8s dashboard version need to be reverted back to v1.8.3 or changed to latest v1.10.0
Affects | Status | Importance | Assigned to | Milestone | ||
---|---|---|---|---|---|---|
Juniper Openstack | Status tracked in Trunk | |||||
R5.0 |
Fix Released
|
Critical
|
Andrey Pavlov | |||
Trunk |
Fix Committed
|
Critical
|
Andrey Pavlov |
Bug Description
Kubernetes dashboard version either need to be reverted back to 1.8.3 version or changed to the latest v1.10.0 as there is no dashboard image available for v1.9.2 as kube version we are using is v1.9.2
https:/
Erlier the dashboard version was hardcoded in contrail anisble deployer which was changes as part of the commit
https:/
[root@nodeg12 contrail-
NAME STATUS ROLES AGE VERSION
nodec60 Ready <none> 2h v1.9.2
nodec61 Ready <none> 2h v1.9.2
nodeg12 NotReady master 2h v1.9.2
[root@nodeg12 logs]# kubectl get pods -n kube-system -o wide
NAME READY STATUS RESTARTS AGE IP NOD
kubernetes-
[root@nodeg12 logs]# kubectl describe pod kubernetes-
Name: kubernetes-
Namespace: kube-system
Node: nodec61/
Start Time: Fri, 19 Oct 2018 22:27:49 +0530
Labels: k8s-app=
Annotations: <none>
Status: Pending
IP: 10.204.217.101
Controlled By: ReplicaSet/
Containers:
kubernetes-
Container ID:
Image: k8s.gcr.
Image ID:
Port: 8443/TCP
Args:
-
State: Waiting
Reason: ImagePullBackOff
Ready: False
Restart Count: 0
Liveness: http-get https:/
Environment: <none>
Mounts:
/certs from kubernetes-
/tmp from tmp-volume (rw)
/
Conditions:
Type Status
Initialized True
Ready False
PodScheduled True
Volumes:
kubernetes-
Type: Secret (a volume populated by a Secret)
SecretName: kubernetes-
Optional: false
tmp-volume:
Type: EmptyDir (a temporary directory that shares a pod's lifetime)
Medium:
kubernetes-
Type: Secret (a volume populated by a Secret)
SecretName: kubernetes-
Optional: false
QoS Class: BestEffort
Node-Selectors: <none>
Tolerations: node-role.
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Pulling 22m (x32 over 2h) kubelet, nodec61 pulling image "k8s.gcr.
Normal BackOff 7m (x661 over 2h) kubelet, nodec61 Back-off pulling image "k8s.gcr.
Warning Failed 2m (x682 over 2h) kubelet, nodec61 Error: ImagePullBackOff
summary: |
- ansible-deployer:k8s dashboard need to be reverted back to v1.8.3 or - changed to latest v1.10.0 + ansible-deployer:k8s dashboard version need to be reverted back to + v1.8.3 or changed to latest v1.10.0 |
workaround is to edit the deployment spec of the dashboard and change the version to v1.8.3 or v1.10.0
root@nodeg12 contrail- test-runs] # kubectl get pods --all-namespaces | grep dashboard dashboard- 5b94954fbf- 24sl7 1/1 ErrImagePull 0 23s
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system kubernetes-
[root@nodeg12 contrail- test-runs] # kubectl edit pod/kubernetes- dashboard- 5b94954fbf- 24sl7 -n kube-system
spec: generate- certificates io/kubernetes- dashboard- amd64:v1. 9.2 io/kubernetes- dashboard- amd64:v1. 10.0
containers:
- args:
- --auto-
image: k8s.gcr.
change to
image: k8s.gcr.
kubectl delete pod kubernetes- dashboard- 5b94954fbf- 24sl7 ...this will create new dashboard pod with the changed version
[root@nodeg12 contrail- test-runs] # kubectl delete pod kubernetes- dashboard- 5b94954fbf- 24sl7 -n kube-system dashboard- 5b94954fbf- 24sl7" deleted test-runs] # kubectl get pods --all-namespaces dashboard- 5b94954fbf- mprxr 0/1 ContainerCreating 0 2s dashboard- 5b94954fbf- 24sl7 0/1 Terminating 2 42m
pod "kubernetes-
[root@nodeg12 contrail-
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system kubernetes-
kube-system kubernetes-
[root@nodeg12 contrail- test-runs] # kubectl get pods --all-namespaces dashboard- 5b94954fbf- mprxr 1/1 Running 0 23s
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system kubernetes-