Skip to content

Conversation

@alinbalutoiu
Copy link

It seems that the cloudflare-tunnel-remote helm chart is missing the --no-autoupdate flag which is added to the default image in the entrypoint looking at the image layer:

ENTRYPOINT ["cloudflared" "--no-autoupdate"]

This results in cloudflared crashing everytime there is a new update. Looking at the logs of the pods stuck in crashbackloopoff:

2023-08-31T19:45:32Z INF cloudflared has been updated version=2023.8.2
2023-08-31T19:45:32Z INF Restarting service managed by SysV...
2023-08-31T19:45:32Z INF PID of the new process is 16
2023-08-31T19:45:32Z ERR Initiating shutdown error="cloudflared has been updated to version 2023.8.2"

It seems that it restarts itself once the update is finished, but since this is a pod, Kubernetes will restart it with the same configuration and cloudflared will be stuck in the same loop forever:

$ kubectl get pods -n cloudflared
NAME                                                   READY   STATUS    RESTARTS      AGE
cloudflared-cloudflare-tunnel-remote-dbcf68f9d-m77s6   1/1     Running   2 (23s ago)   2m

With this PR, the update will be shown in the logs as a warning and the pods will no longer be stuck in crashbackloopoff:

2023-08-31T19:53:08Z WRN Your version 2023.8.1 is outdated. We recommend upgrading it to 2023.8.2

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant