Description of problem:
In RHV-4.2.8-1, 4.2 cluster, with 2 hosts:
1. Host up to date
2. Host with older vdsm (of RHV-4.2.7-9), and a running VM.
This VM is assigned to this 2nd host,
and defined a Manual migration only.
In RHV UI, there is no indication on why the Upgrade host fail.
In engine.log, it is indicated that host can't move to maintenance,
since VMs are not migratable.
Version-Release number of selected component (if applicable):
Update host should fail in such case.
Notification for non migratable vms has been added only in 4.3. Any reason you’re testing it with 4.2.z? In 4.2 this is expected behavior, there is no notification anywhere
There is no intention in testing a 4.3 feature in 4.2.
While I was running Upgrade cluster tests from CFME,
I encountered this RHV bug:
Upgrade host in RHV, in case the host has running VMs,
that are non migratable (manual migration allowed only),
do not fail, as expected.
Clearly the host upgrade is not executed, and the user has no indication in the UI on that.
By the engine.log, it seem RHV gets info on that the VM cannot be migrated.
On RHV-4.2 the engine cannot / do not pass such failure to the Upgrade host role, and this is a known issue?
yes, it never worked in 4.2, until 4.3 manual migration was only for migrations of individual vms.
It only makes sense to test in 4.3
Is there an open bug, targeted to 4.3, for adding migration notification please?
If yes, we can add it here for tracking, and once resolved, this current bug
(of upgrade cluster do not fail on non migratable VMs) can be verified too (on 4.3).
It’s all resolved. Bug 1628484
What was done/is to be verified in this bug?
As after bug 1631215 even manually migrateable VMs are a cause for upgrade to fail (this is expected behaviour now).
Manual move to maintenance of the host before upgrade causes manually migrateable VMs to migrate though, so if this is enough it can be probably verified.
(In reply to Petr Matyáš from comment #6)
> What was done/is to be verified in this bug?
> As after bug 1631215 even manually migrateable VMs are a cause for upgrade
> to fail (this is expected behaviour now).
> Manual move to maintenance of the host before upgrade causes manually
> migrateable VMs to migrate though, so if this is enough it can be probably
Michal can you answer this?
"Manual move to maintenance of the host before upgrade causes manually migrateable VMs to migrate" - yes. That covers also REST API request to move to maintenance. There's no automatic maint.
So lets just verify this bug as the behaviour is as expected.
This bugzilla is included in oVirt 4.3.0 release, published on February 4th 2019.
Since the problem described in this bug report should be
resolved in oVirt 4.3.0 release, it has been closed with a resolution of CURRENT RELEASE.
If the solution does not work for you, please open a new bug report.