Bug 1348950 - [z-stream clone - 3.6.8] Pool VM loses its disk during reinitialisation after shutdown.
Summary: [z-stream clone - 3.6.8] Pool VM loses its disk during reinitialisation after...
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Enterprise Virtualization Manager
Classification: Red Hat
Component: ovirt-engine
Version: 3.6.6
Hardware: Unspecified
OS: Unspecified
unspecified
high
Target Milestone: ovirt-3.6.8
: ---
Assignee: Arik
QA Contact: sefi litmanovich
URL:
Whiteboard:
Depends On: 1346270
Blocks:
TreeView+ depends on / blocked
 
Reported: 2016-06-22 11:53 UTC by rhev-integ
Modified: 2019-12-08 22:49 UTC (History)
12 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Previously, there was a race between the automatic startup of prestarted VMs in a VM pool and manual operations on VMs in the pool. This caused some VMs to lose their disks while being returned to the VM pool. Now, the race has been prevented, so disks are no longer removed when the automatic prestart mechanism is executed in parallel.
Clone Of: 1346270
Environment:
Last Closed: 2016-07-27 14:14:17 UTC
oVirt Team: Virt
Target Upstream Version:
Embargoed:

Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Red Hat Product Errata RHBA-2016:1507 0 normal SHIPPED_LIVE Red Hat Enterprise Virtualization Manager (rhevm) bug fix 3.6.8 2016-07-27 18:10:22 UTC
oVirt gerrit 59450 0 master MERGED core: prevent removal of all images of prestarted stateless vm 2016-06-23 08:55:08 UTC
oVirt gerrit 59468 0 master MERGED core: acquire vm lock on RestoreStatelessVm 2016-06-23 11:49:30 UTC
oVirt gerrit 59697 0 ovirt-engine-3.6 MERGED core: prevent removal of all images of prestarted stateless vm 2016-06-26 07:42:26 UTC
oVirt gerrit 59698 0 ovirt-engine-3.6 MERGED core: acquire vm lock on RestoreStatelessVm 2016-06-26 07:42:21 UTC

Comment 4 sefi litmanovich 2016-07-14 10:59:07 UTC 
I wasn't able to re produce the bug on an env with the same version as reported.
I understand this is a race condition bug and is very tricky to re produce.
Ran whole vm pools test plan on automation, In addition wrote another automated test with a vm pool with all vms set as prestarted, and user allocating vms, then detaching, also allocating and stopping the vms, and a bit of both, all that for around 1 hour.. Did not produce the bug.
I don't think I'll be able to re produce with my resources, seems as though the user had close to 1000 vms on the pool increasing the probability this would happen at some point.

Verified running the same tests on rhevm-3.6.8-0.1.el6.noarch, all tests passed.

With no way to re produce and all vm pools tests passing, I think we can verify this bug, please let me know if there are any objections or suggestions how to test the case if you think this is not sufficient.
Comment 6 errata-xmlrpc 2016-07-27 14:14:17 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHBA-2016-1507.html
Note You need to log in before you can comment on or make changes to this bug.