unable to build: no matches for kind "ClusterIssuer" in version "cert-manager.io/v1alpha2"

[senorsmile]

I'm slowly making my way through the initial deployment of a staging env.

When running:
# helm install -f overrides.yaml djankiserv_test djankiserv/charts/djankiserv/

I get the following error
Error: unable to build kubernetes objects from release manifest: unable to recognize "": no matches for kind "ClusterIssuer" in version "cert-manager.io/v1alpha2"

I confirmed that the cert-manager is installed:

NAME        	NAMESPACE   	REVISION	UPDATED                                	STATUS  	CHART               	APP VERSION
cert-manager	cert-manager	1       	2021-02-20 17:23:43.794226023 +0000 UTC	deployed	cert-manager-v0.16.1	v0.16.1  

Here's my overrides.yaml (with sensitive information scrubbed):

djankiserv:
  host: djankiserv.mydomain.com

ingress:
  annotations:
    cert-manager.io/cluster-issuer: letsencrypt-staging
  tls:
    secretName: letsencrypt-cert-staging

clusterissuer:
  staging:
    enabled: true
    email: myname@gmail.com

[AntonOfTheWoods]

I'll try and run through a full install today. Kubernetes is a strange beast in terms of updates and cert-manager 0.16 is pretty old now. I was going to update to the new 1.2.0. Is there any chance you could update to 1.0.? for testing? Also, have you validated your cert-manager install?

[senorsmile]

I had, although I now notice that one of the deployments is in crash loop:

# kubectl get pods -A
NAMESPACE      NAME                                      READY   STATUS             RESTARTS   AGE
kube-system    hostpath-provisioner-5c65fbdb4f-cdt7s     1/1     Running            0          8h
kube-system    calico-node-8rltk                         1/1     Running            1          8h
kube-system    calico-kube-controllers-847c8c99d-mvq8s   1/1     Running            0          8h
cert-manager   cert-manager-d994d94d7-4vthx              1/1     Running            0          8h
cert-manager   cert-manager-webhook-845d9df8bf-6l5zr     1/1     Running            0          8h
cert-manager   cert-manager-cainjector-fc6c787db-h77hs   0/1     CrashLoopBackOff   96         8h

I'll poke around to see if I can figure out what's up with the failing pod.

[AntonOfTheWoods]

I think it might be related to the cert-manager-cainjector pod. In any case, I just upgraded a 1.20 k3s machine I have to cert-manager 1.2.0 and everything seemed to install pretty flawlessly, provided I changed the clusterissuers to apiVersion: cert-manager.io/v1.

One thing which I'm a little sorry about but will probably not be able to fix without help timewise - I am keeping to the bleeding edge, so unless there is suddenly major interest in the project, will be focusing support on the latest versions of everything, including kubernetes.

I will hopefully be updating the docker image to 0.0.9 today with a new release and will update.

If you think having a clusterissuer instead of an issuer is annoying, I'll definitely consider changing that, though the config is relatively easy if you use other issuers. Let me know.

As an aside, I started out my kube journey on microk8s but have now moved to k3s, as its clustering capabilities were far, far more mature than microk8s when I needed them. It should work everywhere but k3s is likely to be the best supported going forward.

[senorsmile]

I've been messing with k8s for a few years, but have only real played in eks and kubespray for large bare metal stuff. I have also not really used helm (delegating that to people on my teams intsead). So this is both a project to learn stuff I've just been putting off as well as get resilient anki server for myself and a few friends.

I'll continue with the tutorial except upgrading cert-manager to 1.2.0.

I'd be happy to guinea pig any new stuff based on k3s.

[senorsmile]

I am not sure what I had done before, but 1.2.0 cert-manager is up and running. I have a hunch it was the CRD (or lack thereof), but I destroyed that instance and tried fresh. I'll post back and close if I get past the original error.