--delete-access-point-root-dir parameter not working as expected

[rshiva777]

/kind bug

What happened?
I have deployed aws-efs-csi-driver with " - --delete-access-point-root-dir=false" flag to the efs-plugin deployment. So when i create an pvc, access point and pvc-content is created under efs data directory as expected. But when i delete the pvc, access point gets deleted, but the pvc-content is under efs data directory is not getting deleted

What you expected to happen?
pvc-content under efs data directory should be deleted along with pvc and access point

How to reproduce it (as minimally and precisely as possible)?
Install aws-efs-csi-driver with " - --delete-access-point-root-dir=false" flag to the efs-plugin deployment

Anything else we need to know?:

Environment

• Kubernetes version (use kubectl version): 1.25
• Driver version: v1.5.8

Please also attach debug logs to help us better diagnose

• Instructions to gather debug logs can be found here

[rshiva777]

So, as mentioned in the issue #411 i updated the " - --delete-access-point-root-dir=true" flag to the efs-plugin deployment

Now, when i create an pvc, access point and pvc-content is created under efs data directory as expected. But when i delete the pvc, access point is not deleted, and pvc-content is under efs data directory is not getting deleted

pv status is in released state, actually it should be deleted from cluster

when i check the efs-csi-controller logs, found there is some error as below

[alexander-maksymiw-bl]

I had the same issue, IAM policies are according to documentation.
It looks like it's again working when downgrading the aws-efs-csi-driver to e.g. 1.5.4 (the version I was running before)
Could this be related to #1055 or #1001 ?

[RyanStan]

Hi @rshiva777 and @alexander-maksymiw-bl , thank you for bringing this to our attention. Yes, we need to re-add privileged: true to the controller Deployment, since the controller Pods perform a mount to the filesystem to delete the access point directory.

We will push out a new Helm chart release to fix this. And add an integration test to our E2E tests to cover the delete access point root dir option so this doesn't happen again.

[RyanStan]

/close

This has been fixed in the latest release.

[k8s-ci-robot]

@RyanStan: You can't close an active issue/PR unless you authored it or you are a collaborator.

In response to this:

/close

This has been fixed in the latest release.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[mskanth972]

/close

[k8s-ci-robot]

@mskanth972: Closing this issue.

In response to this:

/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.