all pods of the master node become unknown when the apiserver stop

[xiaojingchen]

Bug Report

What version of Kubernetes are you using?

Client Version: version.Info{Major:"1", Minor:"12", GitVersion:"v1.12.2", GitCommit:"17c77c7898218073f14c8d573582e8d2313dc740", GitTreeState:"clean", BuildDate:"2018-10-24T06:54:59Z", GoVersion:"go1.10.4", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"12", GitVersion:"v1.12.2", GitCommit:"17c77c7898218073f14c8d573582e8d2313dc740", GitTreeState:"clean", BuildDate:"2018-10-24T06:43:59Z", GoVersion:"go1.10.4", Compiler:"gc", Platform:"linux/amd64"}

What did you do?

stop a kube-apiserver pod

What did you expect to see?

the k8s cluster is still available

What did you see instead?

[[email protected] kubernetes]# kubectl get po --all-namespaces -o wide |grep 173
default                   stability-proxy-689f47f6bf-95k75                1/1     Unknown     0          13d     10.233.102.168   172.16.4.173   <none>
e2e-cluster1              e2e-cluster1-monitor-7c9cd78fc9-762q6           2/2     Unknown     0          2d      10.233.102.130   172.16.4.173   <none>
e2e-cluster1              e2e-cluster1-tikv-0                             1/1     Unknown     0          4h34m   10.233.102.182   172.16.4.173   <none>
kube-system               calico-node-hclpf                               1/1     NodeLost    8          45d     172.16.4.173     172.16.4.173   <none>
kube-system               coredns-7857b7679f-sdqq2                        1/1     Unknown     0          4h17m   10.233.102.139   172.16.4.173   <none>
kube-system               kube-apiserver-172.16.4.173                     1/1     Unknown     3          14d     172.16.4.173     172.16.4.173   <none>
kube-system               kube-controller-manager-172.16.4.173            1/1     Unknown     15         14d     172.16.4.173     172.16.4.173   <none>
kube-system               kube-proxy-mn6mj                                1/1     NodeLost    2          29d     172.16.4.173     172.16.4.173   <none>
kube-system               kube-scheduler-172.16.4.173                     1/1     Unknown     13         13d     172.16.4.173     172.16.4.173   <none>
kube-system               local-volume-provisioner-snx7r                  1/1     NodeLost    15         45d     10.233.102.162   172.16.4.173   <none>
monitoring                node-exporter-x9zm6                             2/2     NodeLost    16         45d     172.16.4.173     172.16.4.173   <none>
monitoring                prometheus-operator-65c44fb7b7-dbjtf            1/1     Unknown     2          13d     10.233.102.135   172.16.4.173   <none>
pre-stop                  pre-stop-pd-1                                   1/1     Unknown     2          13d     10.233.102.133   172.16.4.173   <none>
stability-cluster1        stability-cluster1-tikv-2                       1/1     Unknown     0          99m     10.233.102.161   172.16.4.173   <none>
stability-cluster2        stability-cluster2-pd-2                         1/1     Unknown     0          98m     10.233.102.169   172.16.4.173   <none>
tidb-cluster              tidb-cluster-pd-1                               1/1     Unknown     0          4h16m   10.233.102.186   172.16.4.173   <none>
tidb-demo                 tidb-demo-discovery-f8f564cb4-zdpfj             1/1     Unknown     0          13d     10.233.102.165   172.16.4.173   <none>
tidb-demo                 tidb-demo-tidb-0                                1/1     Unknown     2595       13d     10.233.102.148   172.16.4.173   <none>
tidb-demo                 tidb-demo-tikv-2                                1/1     Unknown     3620       13d     10.233.102.160   172.16.4.173   <none>
tidb-test                 tidb-test-tikv-3                                1/1     Unknown     0          4h34m   10.233.102.175   172.16.4.173   <none>

[weekface]

@xiaojingchen is working on it.

[aylei]

I suppose we could close this by https://github.com/pingcap/tidb-operator/pull/896/files#diff-8f8eafaeec65b7926966885b1c103ca8R298 ? @weekface

[github-actions]

This issue is stale because it has been open 60 days with no activity. Remove stale label or comment or this will be closed in 15 days