search

gaopeiliang / questions

0 stars 0 forks source link

pod ready when node not ready #7

Open gaopeiliang opened 4 years ago

gaopeiliang commented 4 years ago 

whoami-65c4d6d77-2s6xg                 1/1     Running       0          13m     10.190.12.11     10.190.0.44   <none>           <none>
whoami-65c4d6d77-f6xtt                 1/1     Running       0          13m     10.190.11.10     10.190.0.34   <none>           <none>
gaopeiliang@hk5-test-k8s-master-001:~/test_service$ kubectl get node
NAME          STATUS                        ROLES    AGE   VERSION
10.190.0.27   NotReady                      <none>   32d   v1.14.3
10.190.0.33   Ready                         <none>   32d   v1.14.3
10.190.0.34   NotReady                      <none>   32d   v1.14.3
10.190.0.38   Ready                         <none>   32d   v1.14.3
10.190.0.44   Ready                         <none>   32d   v1.14.3
10.190.0.46   NotReady                      <none>   32d   v1.14.3
10.190.0.47   Ready                         <none>   32d   v1.14.3

在 node 变成 not ready 时，POD的状态 一直处于ready状态，这时service是否该摘除这个endpoint那？ 驱逐中间这段时间，应用可能是无法工作的。。。。

gaopeiliang commented 4 years ago 
16614 event.go:209] Event(v1.ObjectReference{Kind:"Node", Namespace:"", Name:"10.190.0.44", UID:"090efbd8-53d9-11ea-861b-002590eb7f3c", APIVersion:"", ResourceVersion:"", FieldPath:""}): type: 'Normal' reason: 'NodeNotReady' Node 10.190.0.44 status is now: NodeNotReady
Mar 24 09:58:08 hk5-test-k8s-master-001 kube-controller-manager[16614]: I0324 09:58:08.077509   16614 controller_utils.go:140] Updating ready status of pod cfdemo-deployment-5f74f9cb78-58sw9 to false
Mar 24 09:58:08 hk5-test-k8s-master-001 kube-controller-manager[16614]: I0324 09:58:08.093385   16614 controller_utils.go:140] Updating ready status of pod cfdemo-deployment-5f74f9cb78-77nj9 to false
Mar 24 09:58:08 hk5-test-k8s-master-001 kube-controller-manager[16614]: I0324 09:58:08.108280   16614 controller_utils.go:140] Updating ready status of pod out-network-cronjob-1585014600-d8nq7 to false
Mar 24 09:58:08 hk5-test-k8s-master-001 kube-controller-manager[16614]: I0324 09:58:08.116721   16614 controller_utils.go:140] Updating ready status of pod whoami-65c4d6d77-lrbnc to false
Mar 24 09:58:08 hk5-test-k8s-master-001 kube-controller-manager[16614]: I0324 09:58:08.139096   16614 node_lifecycle_controller.go:1059] Controller detected that zone  is now in state PartialDisruption.
Mar 24 09:58:13 hk5-test-k8s-master-001 kube-controller-manager[16614]: I0324 09:58:13.139454   16614 node_lifecycle_controller.go:729] Node 10.190.0.44 is unresponsive as of 2020-03-24 09:58:13.139435547 +0800 CST m=+667774.322163857. Adding it to the Taint queue.

看样子 状态是已经变化了，只是 kubectl get pod 显示的ready状态不对