Created attachment 652827 [details] logs and screen shots Description of problem: create 5 vm's select 3 -> remove -> select the other 2 -> remove -> restart engine. Version-Release number of selected component (if applicable): si24.4 How reproducible: 100% Steps to Reproduce: 1. create 5 vm's 2. select 3 vms -> remove them 3. select 2 vm's -> remove them 4. restart engine Actual results: some of the disks are reported to be changed to illegal state but stay in locked state Expected results: disks should not remain locked. Additional info: logs this also reproduces on scale: Steps to Reproduce: 1. create 100VM's 2. remove all vm's but in bulks of 10-15 vm's (select 15 -> remove -> select 15 -> remove) 3. restart engine.
Dafna, this seems to be like 2 bugs: incorrect reporting in the event log and the disks remaining locked. Can you please separate it to two bugs? Also - do the async tasks remains as zombies, or are they cleared?
*** This bug has been marked as a duplicate of bug 840525 ***
Alon, this bug is about the event log reporting that the disk status has been changed to illegal while they stay in image locked. not related to bug 840525 reopening this bug
The event log should be correct. If disks remain after VM was removed they should move to illegal state, if some of the disks were in other status, then this is a bug.
Already fixed at http://gerrit.ovirt.org/#/c/11075/
Added also a fix which change all the remained disks statuses to illegal with a log message
Patch merged with change id Ic4075c03b8d70d50b69222c73169ccc63e3ab66e
sf5. fixed. disk becomes illegal (not locked) if the engine/vdsm was restarted during vm deletion.
3.2 has been released