helm gitlab/gitlab-omnibus nginx CrashLoopBackOff

I’m trying to install GitLab on a new Kubernetes cluster on Google GCP/GKE. I’m using the the gitlab/gitlab-omnibus helm script as described here https://docs.gitlab.com/ce/install/kubernetes/gitlab_omnibus.html

Almost every things went well except the nginx daemon set that fail with CrashLoopBackOff.

I have used that command
helm install --name teo --set baseDomain=myDomain.com,baseIP=A.B.C.D,legoEmail=myEmail@example.com gitlab/gitlab-omnibus

I’ve removed my sensible information from the command…

Any clue about the problem?

1 Like


i have the exact same issue.

Can anyone help us out here?

@jpblais did you open an issue about this?

No I did not open an issue. If you do, can you give us the link here?

Thank you,

I ran into a crash backoff loop and the culprit was that my application was not responding on the default endpoint for the readinessProbe and livenessProbe. Check your pod logs for these types of errors. I was seeing a 404 not found error, but you may see other errors. Check the helm chart (dig into the hidden default files used by Auto Deploy if you’re using it) for readiness and liveness probe settings.