error: no kind "Installation" is registered for version "operator.tigera.io/v1" in scheme "pkg/scheme/scheme.go:28"

[liweis]

Expected Behavior

pods are running in calico-system namespace after a fresh installation

Current Behavior

no pods in calico-system namespace, and kubectl logs installation.operator.tigera.io/default gets: error: no kind "Installation" is registered for version "operator.tigera.io/v1" in scheme "pkg/scheme/scheme.go:28"

Possible Solution

Steps to Reproduce (for bugs)

1. install k8s 1.30 and calico using https://docs.tigera.io/calico/latest/getting-started/kubernetes/self-managed-onprem/onpremises#install-calico

Context

Your Environment

• Calico version: v3.28.2
• Calico dataplane (iptables, windows etc.)
• Orchestrator version (e.g. kubernetes, mesos, rkt):k8s 1.30
• Operating System and version: Ubuntu 20.04.2 LTS
• Link to your project (optional):

[coutinhop]

@liweis I think what you want to check out are the logs for the tigera operator pod, do kubectl get pods -n tigera-operator to get the pod name, then kubectl logs -n tigera-operator tigera-operator-XXXXX (with the actual pod name you got from the previous command). Could you post the results here, along with the output of kubectl get tigerastatus and maybe also kubectl describe tigerastatus?

[liweis]

@liweis I think what you want to check out are the logs for the tigera operator pod, do kubectl get pods -n tigera-operator to get the pod name, then kubectl logs -n tigera-operator tigera-operator-XXXXX (with the actual pod name you got from the previous command). Could you post the results here, along with the output of kubectl get tigerastatus and maybe also kubectl describe tigerastatus?

@coutinhop Thanks for your reply.

kubectl logs -n tigera-operator tigera-operator-XXXXX gets:

And also:

btw, the kubeadm init cmd I use is: sudo kubeadm init --cri-socket unix:///var/run/cri-dockerd.sock --apiserver-advertise-address=[host ip address] --pod-network-cidr=192.168.0.0/16

[coutinhop]

--apiserver-advertise-address=[host ip address]

could you elaborate? the tigera operator is failing to reach the kubernetes apiserver on 10.96.0.1 from the logs you posted, should it be looking for the apiserver in a different address?

could you also post the output of kubectl get svc -A -o wide ?

[liweis]

--apiserver-advertise-address=[host ip address]

could you elaborate? the tigera operator is failing to reach the kubernetes apiserver on 10.96.0.1 from the logs you posted, should it be looking for the apiserver in a different address?

could you also post the output of kubectl get svc -A -o wide ?

@coutinhop

[liweis]

--apiserver-advertise-address=[host ip address]

could you elaborate? the tigera operator is failing to reach the kubernetes apiserver on 10.96.0.1 from the logs you posted, should it be looking for the apiserver in a different address?

could you also post the output of kubectl get svc -A -o wide ? should I change the calico setup configs to use the api server ip showed here ?

[coutinhop]

@liweis sorry for the delay. In theory, the kubernetes service at 10.96.0.1 should work, so there is something weird going on... I'd say yes, please try pointing the calico configs to the address you have used for the apiserver and let me know how that goes.

Can you post what happens when you use curl on the svc IP (https://10.96.0.1) and on the IP you configured (https://11.165.129.211) from a cluster node?

[mazdakn]

@liweis have you managed to get the information asked above?

[tomastigera]

@liweis is it a regular install or ebpf?

[mazdakn]

Closing due to inactivity. Feel free to re-open with more information asked.