udevadm may be never triggered right now

[ZeroMagic]

In the function VerifyDevicePath(https://github.com/kubernetes-sigs/gcp-compute-persistent-disk-csi-driver/blob/master/pkg/mount-manager/device-utils.go#L89), the value of sdBefore is the same with sdAfter.

Then the function udevadmChangeToDrive will never be called. It seems like the code doesn't actually do anything.

/ping @davidz627

[davidz627]

This is true, the code as-is doesn't do anything :) thanks for the find.

The intent for udevadm triggers is to "workaround" two known errors.

• The disk has attached and shows up as /dev/sdx but does not get symlinked to /dev/disk/by-id/ which we rely on for finding the correct disk
• A disk was detached (at /dev/sdy), a different one has attached also at /dev/sdy and the /dev/disk/by-id/ symlinks are stale (?)
  • @msau42 I feel like I didn't write this second one properly could you confirm/edit

The actual solution to this would be as follows:

• Check that the disk at /dev/disk/by-id/ has a SCSI Serial that matches the diskName that we expect
  • If not, run a udev fix on it
• If the /dev/disk/by-id does not exist, we then have to trigger the udev fix on the /dev/sdx of the missing disk
  • Finding this disk poses an issue, the clumsy method we can think of right now is to take the set of all disks in /dev/disk/by-id and diff them with the set of /dev/sdx and run the udev fix on all the diffs until we get the symlink that we want
    • The issue with this method is that we will trigger udev on disks that may be attached that are not PD repeatedly (this is not ideal but may not be a terrible problem).

@ZeroMagic please let me know and assign yourself if you are interested in working on this fix. If not, considering this is a known issue that occurs infrequently we do not consider this high priority and will fix it sometime before GA.

[ZeroMagic]

@davidz627 Oh, I'm afraid not. I have no time to fix it recently.

[davidz627]

/help

[k8s-ci-robot]

@davidz627:
This request has been marked as needing help from a contributor.

Please ensure the request meets the requirements listed here.

If this request no longer meets these requirements, the label can be removed
by commenting with the /remove-help command.

In response to this:

/help

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.

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[davidz627]

/remove-lifecycle stale

[davidz627]

/assign