I have a docker container that is running fine when I run it using docker run. I am trying to put that container inside a pod but I am facing issues. The first run of the pod shows status as "Completed". And then the pod keeps restarting with CrashLoopBackoff status. The exit code however is 0.
Here is the result of kubectl describe pod :
Name: messagingclientuiui-6bf95598db-5znfh
Namespace: mgmt
Node: db1mgr0deploy01/172.16.32.68
Start Time: Fri, 03 Aug 2018 09:46:20 -0400
Labels: app=messagingclientuiui
pod-template-hash=2695115486
Annotations: <none>
Status: Running
IP: 10.244.0.7
Controlled By: ReplicaSet/messagingclientuiui-6bf95598db
Containers:
messagingclientuiui:
Container ID: docker://a41db3bcb584582e9eacf26b02c7ef26f57c2d43b813f44e4fd1ba63347d3fc3
Image: 172.32.1.4/messagingclientuiui:667-I20180802-0202
Image ID: docker-pullable://172.32.1.4/messagingclientuiui@sha256:89a002448660e25492bed1956cfb8fff447569e80ac8b7f7e0fa4d44e8abee82
Port: 9087/TCP
Host Port: 0/TCP
State: Waiting
Reason: CrashLoopBackOff
Last State: Terminated
Reason: Completed
Exit Code: 0
Started: Fri, 03 Aug 2018 09:50:06 -0400
Finished: Fri, 03 Aug 2018 09:50:16 -0400
Ready: False
Restart Count: 5
Environment Variables from:
mesg-config ConfigMap Optional: false
Environment: <none>
Mounts:
/docker-mount from messuimount (rw)
/var/run/secrets/kubernetes.io/serviceaccount from default-token-2pthw (ro)
Conditions:
Type Status
Initialized True
Ready False
PodScheduled True
Volumes:
messuimount:
Type: HostPath (bare host directory volume)
Path: /mon/monitoring-messui/docker-mount
HostPathType:
default-token-2pthw:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-2pthw
Optional: false
QoS Class: BestEffort
Node-Selectors: <none>
Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s
node.kubernetes.io/unreachable:NoExecute for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 4m default-scheduler Successfully assigned messagingclientuiui-6bf95598db-5znfh to db1mgr0deploy01
Normal SuccessfulMountVolume 4m kubelet, db1mgr0deploy01 MountVolume.SetUp succeeded for volume "messuimount"
Normal SuccessfulMountVolume 4m kubelet, db1mgr0deploy01 MountVolume.SetUp succeeded for volume "default-token-2pthw"
Normal Pulled 2m (x5 over 4m) kubelet, db1mgr0deploy01 Container image "172.32.1.4/messagingclientuiui:667-I20180802-0202" already present on machine
Normal Created 2m (x5 over 4m) kubelet, db1mgr0deploy01 Created container
Normal Started 2m (x5 over 4m) kubelet, db1mgr0deploy01 Started container
Warning BackOff 1m (x8 over 4m) kubelet, db1mgr0deploy01 Back-off restarting failed containerkubectl get pods
NAME READY STATUS RESTARTS AGE
messagingclientuiui-6bf95598db-5znfh 0/1 CrashLoopBackOff 9 23mI am assuming we need a loop to keep the container running in this case. But I dont understand why it worked when it ran using docker and not working when it is inside a pod. Shouldnt it behave the same ?
How do we henerally debug CrashLoopBackOff status apart from running kubectl describe pod and kubectl logs
are you sure you run your software as docker run ... -d ... <command> and it kept running and you use the same exact command in your pod ? In some cases, if you compare things that run on docker with -it and no -d you might find your self in a pinch as they expect terminal to communicate with user and exit if tty is not available (hint: pod/container can be run with tty: true)
It is very unlikely that you have software that runs in a detached docker and does not in kube.
The container would terminate with exit code 0 if there isn't at least one process running in the background. To keep the container running, add these to the deployment configuration:
command: ["sh"]
stdin: trueReplace sh with bash on any other shell that the image may have.
Then you can drop inside the container with exec:
kubectl exec -it <pod-name> shAdd -c <container-name> argument if the pod has more than one container.