-
Notifications
You must be signed in to change notification settings - Fork 275
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
Pods stuck in Pending state while using latest commit #4894
Comments
The rke2 cloud-controller is still running code from 1.26.4, which predates recent dual-stack changes. The following message seen in the cloud-controller-manager logs:
After updating K3s to latest in rancher/image-build-rke2-cloud-provider#30 we are seeing a different error:
This appears to be related to recent changes in dual-stack node-ip behavior on the K3s side - possibly k3s-io/k3s#8581 |
It looks like K3s is handling this by enabling the I also suspect this is a breaking change, as setting the node-ip to a dual-stack list will require the feature-gate to be enabled on ALL cloud-providers used with RKE2, or new nodes will fail to initialize and will remain tainted. This is related to: |
Confirmed that setting the feature-gate on the cloud-controller-manager fixes this (4cb0de3), but I think this should be done on the K3s side, not RKE2. I am also concerned that this will be a breaking change for other cloud providers, unless for some reason the node-ip behavior is gated on use of the built-in stub cloud provider? |
Just confirming that the feature-gate should be enabled for both kubelet AND cloud-controller-manager, starting with 1.27: |
Validated on rke2 version v1.28.3-rc2+rke2r1 (0d0d0e4)Environment DetailsInfrastructure Node(s) CPU architecture, OS, and Version:
Cluster Configuration: Config.yaml:
Steps to reproduce the issue and validate
|
Environmental Info:
RKE2 Version:
rke2 version v1.28.2+dev.45c21222
rke2 version v1.27.6+dev.7c3bb478
Node(s) CPU architecture, OS, and Version:
Ubuntu 22.04, SLES 15 SP4
Cluster Configuration:
Single node
Describe the bug:
Pods are in pending state
Steps To Reproduce:
Install rke2 using commit 45c2122
Expected behavior:
All pods should be scheduled and in running state
Actual behavior:
Pods are waiting to be scheduled
Additional context / logs:
rke2.log
The text was updated successfully, but these errors were encountered: