-
Notifications
You must be signed in to change notification settings - Fork 9
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Kubernetes 1.29 - node NotReady and pods still pending after weave deployment #3
Comments
From the logs, it seems that the weave initialisation procedure has not been able to write a files called To confirm, could you please do the following for me?
{
"cniVersion": "1.0.0",
"name": "weave",
"disableCheck": true,
"plugins": [
{
"name": "weave",
"type": "weave-net",
"hairpinMode": true
},
{
"type": "portmap",
"capabilities": {"portMappings": true},
"snat": true
}
]
}
|
1: kubectl logs -n kube-system weave-net-gftst init The container name seems to be weave-init not init, but no logs, returns empty: 2:
And the file seems to have the same contents you posted:
3: yes, i used kubeadm:
|
Sorry, the container name was indeed Can we check the output of |
|
Yes, weave net is up and running, but the taint is not gone. I think a kubelet restart or a node restart will solve the problem. I currently test weave net with kubernetes versions 1.27 through 1.29, on clusters running debian linux on amd64 and arm64. It works in all those cases. I also use kubeadm, with settings pretty similar to yours. Perhaps I should add coreos to the test mix. |
I tried restarting kubelet and also reboot, and still no go. |
I apologise for inconvenience. I'll try and replicate your environment, and see if I can find the problem. The only thing out of place in your logs is that kubelet has still not detected the CNI setup, even though Weave Net has indicated that the network is available. You can see that in the
|
I tried setting up a cluster on CoreOS, and faced exactly the same problem. But after a long time (over 15 minutes), the node became Ready. This happened with a different CNI as well. |
Yes, seems happens to flannel and weave on clusters above 1.27 on CoreOS, though works fine with calico. |
Also with Openshift OVN-Kubernetes. |
Hi,
I deployed a kubernetes 1.29 cluster and after deploying weave, the node is NotReady and the pods are still in pending.
Same thing with kuberntes 1.28
Last version where it works fine is kubernetes 1.27
Anything else we need to know?
OS: Fedora CoreOS 39
Versions:
Logs:
Any idea what might be wrong?
The text was updated successfully, but these errors were encountered: