Closed otorreno closed 1 year ago
This was already merged into v1.35.0 from another PR: https://github.com/aws/efs-utils/commit/c52de83f3f245b09f5af22e23002995b3d3a5009.
@RyanStan but that was a change to the watchdog folder not to the mount one. The code is duplicated in both places and since such PR you are referencing was merged the code is different
Sorry about that! Thanks for pointing that out. We'll merge this.
@RyanStan do you have an estimate on when this is going to be merged?
@RyanStan when can I expect a new release to be cut for this and the EFS CSI driver? We are facing this problem on a weekly basis and would like to get past it as soon as possible.
Same here - started hitting this issue quite frequently. Would really appreciate a new release with a fix!
@RyanStan could you please release a new version? Thank you.
@RyanStan when would you cut a new release?
Nowadays, the private key generation function checks if the private key file exists. However, if the openssl command that generates the private key file is interrupted before writing the key, the file is there but empty. Such behaviour has been confirmed locally.
That causes the following issue (https://github.com/kubernetes-sigs/aws-efs-csi-driver/issues/683) when the EFS CSI driver pod is restarted. On the next start of the pod in the node, it doesn't generate the private key because of what is stated above.
With this change, if the file is empty it will be generated. No need for the workaround with the initContainer to delete the file.
130 only did it for the watchdog file, we have the same code duplicated in
mount_efs/__init__.py
. Reusing the sameapproach.
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.