Pod ignores drain node command and still on the running state
Issue
-
When draining nodes which run database pods, the database pods are not terminated, even after the grace period has set:
oc adm drain mynode.example.com --grace-period=420 --force --delete-local-data --ignore-daemonsets
-
node log contains lines related to this pod like
<timestamp> <node-fqdn> atomic-openshift-node[17945]: I0624 10:31:10.375739 17945 kuberuntime_container.go:547] Killing container "<docker-container.id>" with 420 second grace period
<timestamp> <node-fqdn> atomic-openshift-node[17945]: W0624 10:31:12.535917 17945 docker_sandbox.go:365] failed to read pod IP from plugin/docker: NetworkPlugin cni failed on the status hook for pod "<pod-id>": Unexpected command output nsenter: cannot open /proc/<id>/ns/net: No such file or directory
<timestamp> <node-fqdn> atomic-openshift-node[17945]: with error: exit status 1
Environment
- Red Hat OpenShift Container Platform (RHOCP) 3.6+
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.