cilium-cli: cilium install fail on k3s ubuntu - Cluster name "" is not valid

Is there an existing issue for this?

  • I have searched the existing issues

What happened?

Can’t install cilium on brand new ubuntu 20.04 with k3s Following this https://docs.cilium.io/en/v1.10/gettingstarted/k3s/

First, with no CNI, node is not ready kubectl get nodes NAME STATUS ROLES AGE VERSION corti2623-awse1pi-1d.corticore.net NotReady control-plane,etcd,master 19m v1.20.15+k3s1

Due to pods stuck in pending state (no CNI) kubectl get pods -A NAMESPACE NAME READY STATUS RESTARTS AGE kube-system coredns-6488c6fcc6-p4crr 0/1 Pending 0 9m31s kube-system local-path-provisioner-5ff76fc89d-lprrr 0/1 Pending 0 9m31s kube-system metrics-server-86cbb8457f-dqth7 0/1 Pending 0 9m31s

kubectl describe pod coredns-6488c6fcc6-p4crr -n kube-system kubectl describe pod local-path-provisioner-5ff76fc89d-lprrr -n kube-system kubectl describe pod metrics-server-86cbb8457f-dqth7 -n kube-system Warning FailedScheduling 9m4s default-scheduler 0/1 nodes are available: 1 node(s) had taint {node.kubernetes.io/not-ready: }, that the pod didn’t tolerate.

Hoping this will be solved when cilium is installed, but that doesn’t work due to Cluster name = “” cilium install ℹ️ using Cilium version “v1.11.1” 🔮 Auto-detected IPAM mode: cluster-pool ❌ Cluster name “” is not valid, must match regular expression: ^a-z0-9$ ↩️ Rolling back installation… Error: Unable to install Cilium: invalid cluster name

Can I set cluster name somewhere?

Cilium Version

v1.11.1 (never made it to installed)

Kernel Version

5.11 Linux corti2623-awse1pi-1d.corticore.net 5.11.0-1020-aws cilium/cilium#21~20.04.2-Ubuntu SMP Fri Oct 1 13:03:59 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

Kubernetes Version

v1.20.15+k3s1 Client Version: version.Info{Major:“1”, Minor:“20”, GitVersion:“v1.20.15+k3s1”, GitCommit:“c7a91e1ccdace64eae4a41b9dc15ef8d7aed6d23”, GitTreeState:“clean”, BuildDate:“2022-01-25T00:24:54Z”, GoVersion:“go1.15.15”, Compiler:“gc”, Platform:“linux/amd64”} Server Version: version.Info{Major:“1”, Minor:“20”, GitVersion:“v1.20.15+k3s1”, GitCommit:“c7a91e1ccdace64eae4a41b9dc15ef8d7aed6d23”, GitTreeState:“clean”, BuildDate:“2022-01-25T00:24:54Z”, GoVersion:“go1.15.15”, Compiler:“gc”, Platform:“linux/amd64”}

Sysdump

cilium never made it to installed

Relevant log output

cilium never made it to installed

Anything else?

No response

Code of Conduct

  • I agree to follow this project’s Code of Conduct

About this issue

  • Original URL
  • State: closed
  • Created 2 years ago
  • Comments: 16 (8 by maintainers)

Commits related to this issue

Most upvoted comments

The docs were changed to mention export KUBECONFIG=/etc/rancher/k3s/k3s.yaml in the k3s-specific guides, see https://docs.cilium.io/en/latest/gettingstarted/k3s/ and https://docs.cilium.io/en/latest/gettingstarted/k8s-install-default/ (k3s tab). As mentioned above, it might take a while for the doc changes to be backported to the v1.10 and v1.11 docs.

I think this should solve the issue originally reported. In case you still see the problem please comment here or open a new issue.

@asloan7 yes, the --disable-network-policy flag should appear in both places. This is already fixed in master branch (see https://docs.cilium.io/en/latest/gettingstarted/k8s-install-default/) and will be fixed in the v1.10 and v1.11 docs once cilium/cilium#18671 is backported to these branches (presumably some time later this week).

@tklauser Better to export KUBECONFIG=/etc/rancher/k3s/k3s.yaml So all subsequent cilium commands work, like cilium status --wait

FWIW, the fact that KUBECONFIG=/etc/rancher/k3s/k3s.yaml is needed for cilium install to work on k3s is documented in the quick start guide, but not the k3s-specific guide. I’ll send a fix for the latter.

EDIT: fix is in second commit of https://github.com/cilium/cilium/pull/18671