[RHOCP 4] Pods are stuck in ContainerCreating state: http2: client connection lost

Solution Verified - Updated -

Issue

  • When deployed pods get stuck in ContainerCreating state. Containers in the pod are stuck in waiting status and the pod never successfully gets an IP assigned.
  • Events from the affected namespace.

    $ oc get events -n <affected-namespace> (openshift-storage in this example)
    
    35s        Warning  FailedCreatePodSandBox  pod/noobaa-operator-xxx      (combined from similar events): Failed to create pod sandbox: rpc error: code = Unknown desc = failed to create pod network sandbox k8s_noobaa-operator-xxx_openshift-storage_4658ed69-8dbb-49ef-b6fc-6575c7b59b43_0(b04981f5be63470994467292c077505bb8b9f78c5379737b5997bbaa415f80de): error adding pod openshift-storage_noobaa-operator-xxx to CNI network "multus-cni-network": Multus: [openshift-storage/noobaa-operator-xxx]: error setting the networks status: SetNetworkStatus: failed to update the pod noobaa-operator-xxx in out of cluster comm: SetNetworkStatus: failed to update the pod noobaa-operator-xxx in out of cluster comm: status update failed for pod openshift-storage/noobaa-operator-xxx: Put "https://[api-int.xxx]:6443/api/v1/namespaces/openshift-storage/pods/noobaa-operator-xxx/status?timeout=1m0s": http2: client connection lost
    
  • If using OVN as the CNI. look at the OVNkube node logs on the affected node. Will repeatedly see this pattern.

    CNI ADD completes
    15:01:09.711165    4536 cni.go:248] [<namespace>/<pod-name>] ADD finished CNI request [<namespace>/<pod-name>], result "{\"interfaces\":[{\"name\":\" <netns_id>\",\"mac\":\"xx:xx:xx:xx:xx:xx\"},{\"name\":\"eth0\",\"mac\":\"xx:xx:xx:xx:xx:xx\",\"sandbox\":\"/var/run/netns/afbab413-4a03-4e82-8877-c61a719ef663\"}],\"ips\":[{\"interface\":1,\"address\":\"xx.xxx.xxx.x/23\",\"gateway\":\"xx.xxx.xxx.x\"}],\"dns\":{}}", err <nil>     
    
    45s later      
    CNI DEL is called:
    15:01:54.908021    4536 cni.go:227][<namespace>/<pod-name>] DEL finished CNI request [<namespace>/<pod-name>], result "{\"dns\":{}}", err <nil>
    
    CNI ADD invoked again                                                                         
    15:01:56.055706    4536 cni.go:227] [<namespace>/<pod-name>] ADD starting CNI request [<namespace>/<pod-name>]       
    

Environment

  • Red Hat OpenShift Container Platform (RHOCP)
    • 4
  • F5 LoadBalancer

Subscriber exclusive content

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.

Current Customers and Partners

Log in for full access

Log In

New to Red Hat?

Learn more about Red Hat subscriptions

Using a Red Hat product through a public cloud?

How to access this content