Closed
Description
What happened:
When volumeAttributes.mountPermissions are set in an NFS PV, we run chmod. The go implementation of os.chmod is such that sometimes it ends up resetting the gid bits. For example, if the mountPermissions are set to "02775" and the NFS target path already has the gid mode bits set, the mode bits from go's perspective change from “20020000775” to “20000000775”.
This causes the CSI driver to unnecessarily apply the SetVolumeOwnership changes later, which takes up quite a lot of time.
What you expected to happen:
The gid bits shouldn't be reset.
How to reproduce it:
Anything else we need to know?:
Environment:
- CSI Driver version:
- Kubernetes version (use
kubectl version
): - OS (e.g. from /etc/os-release):
- Kernel (e.g.
uname -a
): - Install tools:
- Others:
Metadata
Metadata
Assignees
Labels
No labels