Nir Soffer has posted comments on this change.
Change subject: vm.py: State saving in hotunplugDisk. ......................................................................
Patch Set 4:
(3 comments)
https://gerrit.ovirt.org/#/c/45077/4//COMMIT_MSG Commit Message:
Line 9: VM's state was saved before detatching a disk in hotunplugDisk. Line 10: Line 11: Saving vm's state should be performed only AFTER vdsm is actually detaching Line 12: the disk, otherwise the disk will remain in xmlDesc and be calculated in Line 13: the vm's hash. Saving state does not effect xmlDesc. Line 14: This causes a bug in the engine's VM monitoring: the engine Line 15: gets the same hash as before the disk was unplugged, thus does not Line 16: update the disk's status. Line 17:
Line 12: the disk, otherwise the disk will remain in xmlDesc and be calculated in Line 13: the vm's hash. Line 14: This causes a bug in the engine's VM monitoring: the engine Line 15: gets the same hash as before the disk was unplugged, thus does not Line 16: update the disk's status. How saving state before removal can cause engine to get same hash as before the removal? Line 17: Line 18: Change-Id: I2cf18186cbba33d7e74fd15651ffec3149c98e1d Line 19: Bug-Url:https://bugzilla.redhat.com/1206696 Line 20: Bug-Url: https://bugzilla.redhat.com/1044466
Line 16: update the disk's status. Line 17: Line 18: Change-Id: I2cf18186cbba33d7e74fd15651ffec3149c98e1d Line 19: Bug-Url:https://bugzilla.redhat.com/1206696 Line 20: Bug-Url: https://bugzilla.redhat.com/1044466 This change looks good, but I don't think it is required for fixing this bug.
Unless you can explain how having incorrect state on disk effect these bugs.