Back to bug 1389332

Who When What Removed Added
Red Hat Bugzilla Rules Engine 2016-10-27 13:08:50 UTC Depends On 1341106
Target Milestone ovirt-4.0.6 ---
meital avital 2016-11-02 12:32:19 UTC Target Milestone --- ovirt-4.0.6
Link ID oVirt gerrit 65394
Link ID oVirt gerrit 64991
Link ID oVirt gerrit 65342
Link ID oVirt gerrit 64994
QA Contact mavital alukiano
Vinzenz Feenstra [evilissimo] 2016-11-03 12:04:08 UTC Status NEW MODIFIED
Vinzenz Feenstra [evilissimo] 2016-11-07 08:57:05 UTC Doc Text Cause:
When VMs where stopped during a system shutdown/reboot the shutdown of those instances looked to VDSM the same as if the VM would have been shutdown gracefully from within the Guest Operating System. In this case we have sent this information to the RHV Engine.

Consequence:
The RHV Engine did not start HA VMs on a different Host because it considered the stopping of the VMs as user initiated.

Fix:
With help of the RHV guest agent, vdsm will now detect that a VM was shutdown from within the system and therefore can differentiate the unplanned shutdown and reports this information from now on accordingly

Result:

HA VMs stopped on a system shutdown (e.g. initiated through fencing) are now restarted on a different system.
Doc Type If docs needed, set a value Bug Fix
Artyom 2016-11-21 14:41:35 UTC Status MODIFIED ON_QA
Keywords Triaged
Status ON_QA VERIFIED
Michal Skrivanek 2016-12-12 12:51:13 UTC Status VERIFIED MODIFIED
Michal Skrivanek 2016-12-12 12:52:57 UTC Target Milestone ovirt-4.0.6 ovirt-4.0.7
Summary [z-stream clone - 4.0.6] [RHEV 3.6.5] HA vms do not start after successful power-management. [z-stream clone - 4.0.7] [RHEV 3.6.5] HA vms do not start after successful power-management.
Megan Lewis 2016-12-13 10:10:31 UTC CC melewis
Doc Text Cause:
When VMs where stopped during a system shutdown/reboot the shutdown of those instances looked to VDSM the same as if the VM would have been shutdown gracefully from within the Guest Operating System. In this case we have sent this information to the RHV Engine.

Consequence:
The RHV Engine did not start HA VMs on a different Host because it considered the stopping of the VMs as user initiated.

Fix:
With help of the RHV guest agent, vdsm will now detect that a VM was shutdown from within the system and therefore can differentiate the unplanned shutdown and reports this information from now on accordingly

Result:

HA VMs stopped on a system shutdown (e.g. initiated through fencing) are now restarted on a different system.
Previously, when virtual machines were stopped during a system shutdown or reboot they appeared to VDSM as if they had been shut down from within the guest operating system. This information was sent to the Red Hat Virtualization Manager and this meant that the Manager did not start highly available virtual machines on a different host because the Manager considered the the shut down of the virtual machines to be user-initiated. Now, the Red Hat Virtualization guest agent allows VDSM to detect that the virtual machine was shut down from within the system and this means that highly available virtual machines stopped during a system shutdown are restarted on a different host.
errata-xmlrpc 2016-12-19 13:57:31 UTC Status MODIFIED ON_QA
Artyom 2016-12-19 15:52:47 UTC Flags needinfo?(michal.skrivanek)
Michal Skrivanek 2016-12-19 15:56:36 UTC Summary [z-stream clone - 4.0.7] [RHEV 3.6.5] HA vms do not start after successful power-management. [el5][el6][el7-ppc64][el7-ppc64le] HA vms do not start after successful power-management.
Flags needinfo?(michal.skrivanek)
Artyom 2016-12-21 10:27:21 UTC Status ON_QA ASSIGNED
Artyom 2016-12-21 11:31:35 UTC Status ASSIGNED MODIFIED
errata-xmlrpc 2017-01-02 14:44:48 UTC Status MODIFIED ON_QA
Artyom 2017-01-03 16:49:59 UTC Status ON_QA VERIFIED
Hanns-Joachim Uhl 2017-02-05 16:13:34 UTC CC hannsj_uhl
Blocks 1359843
errata-xmlrpc 2017-03-15 13:12:04 UTC Status VERIFIED RELEASE_PENDING
errata-xmlrpc 2017-03-16 15:28:02 UTC Status RELEASE_PENDING CLOSED
Resolution --- ERRATA
Last Closed 2017-03-16 11:28:02 UTC

Back to bug 1389332