https://rancher.com/ logo
Title
m

mysterious-wire-57288

01/11/2023, 1:05 PM
I believe rancher and k3s was developed by same organizations but rancher installation with helm isn't working because of kubeversion specification on rancher helm chart. It requires version to be v1.25.4 but actual node version using k3s is v1.25.4+k3s1. What is the fix here?
c

creamy-pencil-82913

01/11/2023, 4:38 PM
use a compatible version of K3s?
We don’t wait to release new versions of K3s until Rancher supports them….
Rancher is a complicated product and some times it takes a couple months before it supports new Kubernetes minor releases.
Also I think you’re misreading the error from the chart. I suspect it is probably telling you to use 1.24.x
m

mysterious-wire-57288

01/11/2023, 6:41 PM
I'm using v1.25.4+k3s1. Chart compatible is v1.25.4. I think extra k3s1 is the issue.
helm thinks its a different version.
apiVersion: <http://argoproj.io/v1alpha1|argoproj.io/v1alpha1>
kind: Application
metadata: 
  name: rancher
  namespace: argocd

  finalizers:
    - <http://resources-finalizer.argocd.argoproj.io|resources-finalizer.argocd.argoproj.io>
spec:
  project: default
  source:
    repoURL: '<https://releases.rancher.com/server-charts/stable>'
    targetRevision: 2.7.0
    helm:
      parameters:
        - name: hostname
          value: "<http://rancher.example.com|rancher.example.com>"
        - name: ingress.tls.source
          value: "letsEncrypt"
        - name: letsEncrypt.email
          value: "<mailto:example@gmail.com|example@gmail.com>"
        - name: letsEncrypt.ingress.class
          value: "traefik"
        - name: replicas
          value: "3"
         
    chart: rancher
  destination:
    server: '<https://kubernetes.default.svc>'
    namespace: cattle-system
  syncPolicy:
    automated:
      prune: true
      selfHeal: true
    syncOptions:
      - CreateNamespace=true
It was the argocd manifest I have used.
It doesn't work, unless I add extra +k3s1 version on kubeVersion.
c

creamy-pencil-82913

01/11/2023, 6:59 PM
the kubeVersion where?
we are constantly deploying Rancher to k3s clusters without issue
I suspect that this is an issue with argo, not the chart itself. have you confirmed whether or not this same problem occurs if you use the helm cli to deploy the chart?
m

mysterious-wire-57288

01/11/2023, 7:01 PM
Same problem with helm cli.
& as @careful-piano-35019 said that might be the issue too.
c

creamy-pencil-82913

01/11/2023, 7:04 PM
The error you are getting is this right:
brandond@dev01:~$ helm install rancher rancher-server-charts/rancher --version=2.7.0
Error: INSTALLATION FAILED: chart requires kubeVersion: < 1.25.0-0 which is incompatible with Kubernetes v1.25.4+k3s1
that says you need 1.24.x. Not 1.25.4 instead of 1.25.4
Also I think you’re misreading the error from the chart. I suspect it is probably telling you that 1.25.4 (the version you’re using) is not supported
requires kubeVersion: < 1.25.0-0
1.25.4 is not less than 1.25.0
m

mysterious-wire-57288

01/11/2023, 7:06 PM
Now got it. 😅
I thought problem was +k3s1
c

creamy-pencil-82913

01/11/2023, 7:06 PM
no
the problem is that you’re on 1.25
m

mysterious-wire-57288

01/11/2023, 7:07 PM
so what I did is overwrite that chart and deployed.
c

creamy-pencil-82913

01/11/2023, 7:07 PM
which is what Paul and I have been saying
m

mysterious-wire-57288

01/11/2023, 7:07 PM
Then all messed up
c

creamy-pencil-82913

01/11/2023, 7:07 PM
don’t do that
install k3s 1.24
you’ll have to start over with a new cluster, as you can’t downgrade from one minor to another
m

mysterious-wire-57288

01/11/2023, 7:08 PM
I'm sorry for wasting time🙌
I think at the moment, I have my cluster with lots of thing running. Hard for me to downgrade version because I don't know how well other components will behave. I'll wait for new release from rancher.
I'm still facing that metrics server issue as discussed before.
c

creamy-pencil-82913

01/11/2023, 7:10 PM
we don’t technically support running rancher alongside other workloads. the rancher local cluster should only be used to manage downstream clusters. We won’t stop you from doing that though.
m

mysterious-wire-57288

01/11/2023, 7:11 PM
My cluster is solely made for gitops type deployments. Argocd is way to go for us. But I've seen a lots of video on rancher, So thought of trying that out too.
But rancher has never worked for me. I have had few attempts before. Maybe always that compatibility issue.