Support on investigating an issue with JuiceFS on EKS

[nhannn-imt]

Hello everyone,
I’ve had successful deployments with JuiceFS on other EKS clusters, but this time I’m stuck at a step that’s confusing me. I deployed the CSI driver to my new EKS cluster (EKS managed node groups) using the Helm chart, attached the necessary IAM policy to the cluster to access S3, then deployed a test pod to a namespace to verify connectivity to S3. I’m seeing an issue: the pod hangs at juicefs[7] <INFO>: Data use s3://juicefs-cc-prod/data/ [format.go:484] and is terminated after 60 seconds due to a timeout.

Normally, if the cluster lacks access to S3, I’d expect a permissions error, but in this case it just loops—terminates on timeout and restarts—so the pod never becomes Ready.

I would greatly appreciate any advice or suggestions—thank you in advance for your help!

[zxh326]

Can you turn on debug logs and check again? Specifically, you can add debug to the mount options.

[nhannn-imt]

I just ran the command again with --verbose to see the logs, and it works :( I haven't changed anything. But it works nevertheless, so thank you