This message was deleted.

You might be having the same issue as this? https://github.com/rancher/fleet/issues/2053#issuecomment-1896067091

e65729aaee7f:/var/lib/rancher # kubectl -n fleet-default get secret custom-199db9026ce4-machine-plan                         
NAME                               TYPE                         DATA   AGE                                                   
custom-199db9026ce4-machine-plan   <http://rke.cattle.io/machine-plan|rke.cattle.io/machine-plan>   13     370d

Not actually. I exec the command in my Rancher(Docker Container)

Thanks, the new machine/node must be the one ending

f95cb

? is the secret for that one empty?

Yes.

custom-50af4a3f95cb-machine-plan

is 23.hdp.dc.xyz.

is the secret for that empty?

The new node

custom-50af4a3f95cb-machine-plan

is "readd" node. It is K8S Node (Role: ALL) as yestoday occurs message like (Waiting for probes: kube-controller-manager, kube-scheduler). Then I delete the node and I clean it clear deeply.(The node must be without any rke2/rancher/k3s stuff things.).

kubectl -n fleet-default get secret custom-50af4a3f95cb-machine-plan

shows an empty Secret?

YES.

e65729aaee7f:/var/lib/rancher # kubectl -n fleet-default get secret custom-50af4a3f95cb-machine-plan                         
NAME                               TYPE                         DATA   AGE                                                   
custom-50af4a3f95cb-machine-plan   <http://rke.cattle.io/machine-plan|rke.cattle.io/machine-plan>   0      155m

BTW.

ok, I'm kinda glad (sorry) to see that as it means we're not the only people experiencing the issue. I've not found a solution yet unfortunately, it's stopped us registering nodes to new clusters on two Rancher clusters.

You mean it's the same issue as you pasted?https://github.com/rancher/fleet/issues/2053

I think so yes. I raised that as a fleet issue to start with as it seemed to be related to an issue in the fleet logs but I don't think it is fleet. the machine-plan secret isn't getting populated so rancher-system-agent on the node is endlessly polling for a non-existent plan.

fleet-agent where is it ?

e65729aaee7f:/var/lib/rancher # kubectl get ns  | grep fleet                                                                 
cattle-fleet-clusters-system                Active        539d                                                               
cattle-fleet-local-system                   Active        539d                                                               
cattle-fleet-system                         Active        539d                                                               
cluster-fleet-default-jarvis-4c73dcb0c94c   Active        539d                                                               
cluster-fleet-local-local-1a3d67d0a899      Active        539d                                                               
fleet-default                               Active        539d                                                               
fleet-local                                 Active        539d

I've not found any logs that help locate the issue unfortunately.

@prehistoric-gpu-28362 https://github.com/rancher/rancher/issues/41125 We solved the problem as the comment.

Thanks @busy-country-65143