fail to mark edge-node

[GitHubThinking]

What happened:

The edge-node is not marked.

[root@k8smaster bin]# ./yurtctl convert --provider kubeadm --cloud-nodes k8smaster

I1125 10:47:21.752654   17882 convert.go:318] mark k8smaster as the cloud-node
I1125 10:48:01.793163   17882 util.go:542] servant job(yurtctl-disable-node-controller-k8smaster) has succeeded
I1125 10:48:01.793210   17882 convert.go:343] complete disabling node-controller
I1125 10:48:01.794493   17882 convert.go:443] kube-public/cluster-info configmap already exists, skip to prepare it
I1125 10:48:01.805102   17882 convert.go:408] deploying the yurt-hub and resetting the kubelet service on edge nodes...
E1125 10:48:01.808416   17882 util.go:539] fail to run servant job(yurtctl-servant-convert-k8snode2): jobs.batch "yurtctl-servant-convert-k8snode2" already exists
E1125 10:48:01.808802   17882 util.go:539] fail to run servant job(yurtctl-servant-convert-k8snode1): jobs.batch "yurtctl-servant-convert-k8snode1" already exists
I1125 10:48:01.808812   17882 convert.go:414] complete deploying yurt-hub on edge nodes
I1125 10:48:01.808817   17882 convert.go:417] deploying the yurt-hub and resetting the kubelet service on cloud nodes
E1125 10:48:01.811134   17882 util.go:539] fail to run servant job(yurtctl-servant-convert-k8smaster): jobs.batch "yurtctl-servant-convert-k8smaster" already exists
I1125 10:48:01.811145   17882 convert.go:423] complete deploying yurt-hub on cloud nodes

[root@k8smaster bin]# ./yurtctl markautonomous
W1125 10:57:39.404158   20585 markautonomous.go:118] there is no edge nodes, please label the edge node first

In log,it is showed that

11月 25 10:57:51 k8smaster kubelet[1660]: E1125 10:22:51.583366    1660 pod_workers.go:191] Error syncing pod 740f5710-694d-4c13-b251-1d8797403b17 ("yurtctl-servant-revert-k8smaster-4pgh6_kube-system(740f5710-694d-4c13-b251-1d8797403b17)"), skipping: failed to "StartContainer" for "yurtctl-servant" with CrashLoopBackOff: "back-off 2m40s restarting failed container=yurtctl-servant pod=yurtctl-servant-revert-k8smaster-4pgh6_kube-system(740f5710-694d-4c13-b251-1d8797403b17)"

What you expected to happen:

I want to see the result:
'mark k8smaster as the edge-node'

and "markautonomous" is successful.

How to reproduce it (as minimally and precisely as possible):

Anything else we need to know?:

Environment:

• OpenYurt version: v0.5.0
• Kubernetes version (use kubectl version): v1.18.0
• OS (e.g: cat /etc/os-release): centos7
• Kernel (e.g. uname -a):
• Install tools:
• Others:

others

/kind bug

[rambohe-ch]

@GitHubThinking Thank you for raising issue.

/assign @adamzhoul @Peeknut

[adamzhoul]

E1125 10:48:01.808416   17882 util.go:539] fail to run servant job(yurtctl-servant-convert-k8snode2): jobs.batch "yurtctl-servant-convert-k8snode2" already exists
E1125 10:48:01.808802   17882 util.go:539] fail to run servant job(yurtctl-servant-convert-k8snode1): jobs.batch "yurtctl-servant-convert-k8snode1" already exists
E1125 10:48:01.811134   17882 util.go:539] fail to run servant job(yurtctl-servant-convert-k8smaster): jobs.batch "yurtctl-servant-convert-k8smaster" already exists

maybe the old job is stopping yurtctl convert label the node.
please delete the job, and yurtctl reset then yurtctl convert to do all those again.

issue alike #572

[GitHubThinking]

E1125 10:48:01.808416   17882 util.go:539] fail to run servant job(yurtctl-servant-convert-k8snode2): jobs.batch "yurtctl-servant-convert-k8snode2" already exists
E1125 10:48:01.808802   17882 util.go:539] fail to run servant job(yurtctl-servant-convert-k8snode1): jobs.batch "yurtctl-servant-convert-k8snode1" already exists
E1125 10:48:01.811134   17882 util.go:539] fail to run servant job(yurtctl-servant-convert-k8smaster): jobs.batch "yurtctl-servant-convert-k8smaster" already exists

maybe the old job is stopping yurtctl convert label the node. please delete the job, and yurtctl reset then yurtctl convert to do all those again.

issue alike #572

Please tell me how to delete the old job, I'm a beginner and hope you can instruct me step by step.

[Peeknut]

You can get the job in ns: kube-system, and you will see the convert jobs.

kubectl get job -n kube-system -A

then you can delete the specified job by its namespace and name with

kubectl delete job -n < ns> < job-name>

here you can run:

kubectl delete job -n kube-system yurtctl-servant-convert-k8snode2
kubectl delete job -n kube-system yurtctl-servant-convert-k8snode1
kubectl delete job -n kube-system yurtctl-servant-convert-k8smaster

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.