xgboost-operator
xgboost-operator copied to clipboard
Cannot start xgboost-operator container
I am trying to deploy operator by kustomize. But looks like the pod is not running, and the status is CrashLoopBackOff. Then look into detail by using describe command, I have following message: Last State: Terminated Reason: ContainerCannotRun Message: OCI runtime create failed: container_linux.go:349: starting container process caused "exec: "/root/manager": stat /root/manager: no such file or directory": unknown I did some research online, but didn't get the answer I want. Some people said this is a Windows specific problem, just use kubectl exec -it [pod_name] -c [container_name] -- ./bin/sh. But the thing is the container is failed to be created, so cannot use kubectl exec -it command. Kind of confused. Appreciate any help
Issue-Label Bot is automatically applying the labels:
| Label | Probability |
|---|---|
| area/operator | 0.60 |
| kind/bug | 0.57 |
Please mark this comment with :thumbsup: or :thumbsdown: to give our bot feedback! Links: app homepage, dashboard and code for this bot.
@merlintang @terrytangyuan
Im getting the same issue, any ideas @merlintang @terrytangyuan ?
Seems like the issue might be that /root/manager was not built successfully or not copied successfully to the next phase https://github.com/kubeflow/xgboost-operator/blob/78f8cf50bb943247e038a8feb5a9f7e47d810d65/Dockerfile#L19. Are you able to build the Docker image successfully?
@terrytangyuan I tried to use docker to build the image, but the same problem appear. I build by using our latest image in our repository by pull and start command. Still not sure is this issue due to my environment setup or xgboost-operator image. Cause I don't have this issue before
Can you show how I can deploy xgboost operator via building docker image and running it vs using the kustomize manifests as shown in the official readme? I'm also getting the same error. I'm deploying it alongside my kubeflow environment in vanilla k8s.
@xfate123 can you check the updated version work or not ?
@merlintang sure,my pleasure