Azure AKS: No matches for kind "CustomResourceDefinition" in version "apiextensions.k8s.io/v1beta1"

[Sdelausnay]

Discussed in #980

^{Originally posted by jhall-bondi September 1, 2022}
Hi


Following the quick start guide setting up the project in WSL. Came across this error after using the "opctl apply" which has me stumped.


output below:

opctl apply
Starting deployment...

clusterrole.rbac.authorization.k8s.io/application-manager-role configured
clusterrolebinding.rbac.authorization.k8s.io/application-manager-rolebinding configured
namespace/application-system unchanged
service/application-controller-manager-service unchanged
statefulset.apps/application-controller-manager unchanged

Failed: unable to recognize ".onepanel/application.kubernetes.yaml": no matches for kind "CustomResourceDefinition" in version "apiextensions.k8s.io/v1beta1"

Some googling found this issue. https://stackoverflow.com/questions/69054622/unable-to-install-crds-in-kubernetes-kind
So either I have not followed a step correctly, or maybe the manifest is applying outdated API versions?


Some pointers in the right direction would be appreciated.



Josh.

[hsgBlueSmoke]

I got the same error on AWS EKS.

opctl apply gives this output

`Starting deployment...

clusterrole.rbac.authorization.k8s.io/application-manager-role configured
clusterrolebinding.rbac.authorization.k8s.io/application-manager-rolebinding configured
namespace/application-system unchanged
service/application-controller-manager-service unchanged
statefulset.apps/application-controller-manager unchanged

Failed: unable to recognize ".onepanel/application.kubernetes.yaml": no matches for kind "CustomResourceDefinition" in version "apiextensions.k8s.io/v1beta1"`

opctl app status gives me

`Your deployment is NOT ready; not all Pods are running. To view all Pods:
$ kubectl get pods -A
error: services "istio-ingressgateway" not found
In your DNS, add a CNAME record for *** and point it to
Once complete, your application will be running at http://app.***

ArtifactRepository tests: Successful`

ectl get pods -A gives the output

NAMESPACE NAME READY STATUS RESTARTS AGE application-system application-controller-manager-0 1/1 Running 0 7h14m kube-system aws-node-n6rnk 1/1 Running 0 119m kube-system aws-node-rnlft 1/1 Running 1 (7h18m ago) 7h20m kube-system coredns-79cbbf8cb-ht6hg 1/1 Running 0 119m kube-system coredns-79cbbf8cb-lf59m 1/1 Running 0 119m kube-system kube-proxy-8f77k 1/1 Running 0 7h20m kube-system kube-proxy-9d8b9 1/1 Running 0 119m

I can not get rid of the feeling that it is a dns error. but do not come on it.

[kchudy-krx]

I have the same error on AKS.
After digging some documentation, I see that apiextensions.k8s.io/v1beta1 is no longer served as of v1.22.

K8s deprecation

Right now, AKS runs on higher k8s version (1.23.8 is the lowest, 1.24.6 default).

[Zeleni9]

Do you know if it is possible to deploy it on Azure when the server version can not be lower than 1.25, even though I get the client on 1.21.x to make it work with v1beta1.

I even tried changing the file definition to be v1 and change .onepanel/application.kubernetes.yaml to be updated version but there are a lot of files overwritting it to the v1beta1 version and now it cannot be deployed.