systemd stops reading and processing dbus event of runc cgroup invokes
Issue
- Pods which are started on a node are stuck in the state
ContainerCreating
until they are manually deleted or the host is rebooted. - Pods are stuck in
Terminating
state until the OpenShift Node is manually rebooted. - There were problems with deployments because most deploy pods have been in
ContainerCreating
state for more than a day. - Starting a container using
docker
fails with/usr/bin/docker-current: Error response from daemon: containerd: container did not start before the specified timeout
- We are seeing a ton of the below messages in
journal
and are unable to start any new container usingdocker
on this particular node.
Jun 24 10:10:26 node123 crond[111309]: pam_systemd(crond:session): Failed to create session: Connection timed out
Jun 24 10:10:26 node123 systemd-logind[10714]: Failed to start user slice user-0.slice, ignoring: Connection timed out ((null))
Jun 24 10:10:51 node123 systemd-logind[10714]: Failed to start session scope session-13692.scope: Connection timed out
- Reporting a strange error message from
systemd
after which various operation related toDBus
are failing:
Jun 04 14:03:35 node123 systemd[1]: Failed to propagate agent release message: Operation not supported
- Containers stuck in ContainerCreating and below messages are see in the journal logs.
Jan 01 01:01:01 hostname atomic-openshift-node: I0530 01:01:01.145075 91428 server.go:470] type: 'Warning' reason: 'FailedCreatePodSandBox' Failed create pod sandbox: rpc error: code = Unknown desc = failed to start sandbox container for pod "pod-name": Error response from daemon: oci runtime error: The maximum number of active connections for UID 0 has been reached
Environment
- Red Hat Enterprise Linux 7
- Openshift Container Platform 3.x
- systemd
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.