Using helm for deploying chart on my Kubernetes cluster, since one day, I can't deploy a new one or upgrading one existed.
Indeed, each time I am using helm I have an error message telling me that it is not possible to install or upgrade ressources.
If I run helm install --name foo . -f values.yaml --namespace foo-namespace I have this output:
Error: release foo failed: the server could not find the requested resource
If I run helm upgrade --install foo . -f values.yaml --namespace foo-namespace or helm upgrade foo . -f values.yaml --namespace foo-namespace I have this error:
Error: UPGRADE FAILED: "foo" has no deployed releases
I don't really understand why.
This is my helm version:
Client: &version.Version{SemVer:"v2.12.3", GitCommit:"eecf22f77df5f65c823aacd2dbd30ae6c65f186e", GitTreeState:"clean"}
Server: &version.Version{SemVer:"v2.12.3", GitCommit:"eecf22f77df5f65c823aacd2dbd30ae6c65f186e", GitTreeState:"clean"}
On my kubernetes cluster I have tiller deployed with the same version, when I run kubectl describe pods tiller-deploy-84b... -n kube-system:
Name: tiller-deploy-84b8...
Namespace: kube-system
Priority: 0
PriorityClassName: <none>
Node: k8s-worker-1/167.114.249.216
Start Time: Tue, 26 Feb 2019 10:50:21 +0100
Labels: app=helm
name=tiller
pod-template-hash=84b...
Annotations: <none>
Status: Running
IP: <IP_NUMBER>
Controlled By: ReplicaSet/tiller-deploy-84b8...
Containers:
tiller:
Container ID: docker://0302f9957d5d83db22...
Image: gcr.io/kubernetes-helm/tiller:v2.12.3
Image ID: docker-pullable://gcr.io/kubernetes-helm/tiller@sha256:cab750b402d24d...
Ports: 44134/TCP, 44135/TCP
Host Ports: 0/TCP, 0/TCP
State: Running
Started: Tue, 26 Feb 2019 10:50:28 +0100
Ready: True
Restart Count: 0
Liveness: http-get http://:44135/liveness delay=1s timeout=1s period=10s #success=1 #failure=3
Readiness: http-get http://:44135/readiness delay=1s timeout=1s period=10s #success=1 #failure=3
Environment:
TILLER_NAMESPACE: kube-system
TILLER_HISTORY_MAX: 0
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from helm-token-... (ro)
Conditions:
Type Status
Initialized True
Ready True
ContainersReady True
PodScheduled True
Volumes:
helm-token-...:
Type: Secret (a volume populated by a Secret)
SecretName: helm-token-...
Optional: false
QoS Class: BestEffort
Node-Selectors: <none>
Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s
node.kubernetes.io/unreachable:NoExecute for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 26m default-scheduler Successfully assigned kube-system/tiller-deploy-84b86cbc59-kxjqv to worker-1
Normal Pulling 26m kubelet, k8s-worker-1 pulling image "gcr.io/kubernetes-helm/tiller:v2.12.3"
Normal Pulled 26m kubelet, k8s-worker-1 Successfully pulled image "gcr.io/kubernetes-helm/tiller:v2.12.3"
Normal Created 26m kubelet, k8s-worker-1 Created container
Normal Started 26m kubelet, k8s-worker-1 Started container
Is someone have faced the same issue ?
Update:
This the folder structure of my actual chart named foo: structure folder of the chart:
> templates/
> deployment.yaml
> ingress.yaml
> service.yaml
> .helmignore
> Chart.yaml
> values.yaml
I have already tried to delete the chart in failure using the delete command helm del --purge foo but the same errors occurred.
Just to be more precise, the chart foo is in fact a custom chart using my own private registry. ImagePullSecret are normally setting up.
I have run these two commands helm upgrade foo . -f values.yaml --namespace foo-namespace --force | helm upgrade --install foo . -f values.yaml --namespace foo-namespace --force and I still get an error:
UPGRADE FAILED
ROLLING BACK
Error: failed to create resource: the server could not find the requested resource
Error: UPGRADE FAILED: failed to create resource: the server could not find the requested resource
Notice that foo-namespace already exist. So the error don't come from the namespace name or the namespace itself. Indeed, if I run helm list, I can see that the foo chart is in a FAILED status.
Tiller stores all releases as ConfigMaps in Tiller's namespace(kube-system in your case). Try to find broken release and delete it's ConfigMap using commands:
$ kubectl get cm --all-namespaces -l OWNER=TILLER
NAMESPACE NAME DATA AGE
kube-system nginx-ingress.v1 1 22h
$ kubectl delete cm nginx-ingress.v1 -n kube-system
Next, delete all release objects (deployment,services,ingress, etc) manually and reinstall release using helm again.
If it didn't help, you may try to download newer release of Helm (v2.14.3 at the moment) and update/reinstall Tiller.
I had the same issue, but cleanup did not help also try the same helm chart on a brand new k8s cluster did not help.
So I found out that there was a missing apiVersion caused the problem. I found it out by doing a
helm install xyz --dry-run
copy the output to a new test.yaml file and use
kubectl apply test.yaml
there I see the error (apiVersion line was moved to a comment line)
I had the same problem but not due to broken releases. After upgrading helm. It seems newer versions of helm do bad with the --wait parameter. So for anyone facing the same issue: Just removing --wait, and leave --debug from helm upgrade parameters solved my issue.
来源：https://stackoverflow.com/questions/54883258/helm-install-or-upgrade-release-failed-on-kubernetes-cluster-the-server-could-n