Bug 1035289

Summary: comment to "restart vm after enabling virtIO-SCSI" in GUI
Product: Red Hat Enterprise Virtualization Manager Reporter: Aharon Canan <acanan>
Component: ovirt-engine-webadmin-portalAssignee: Daniel Erez <derez>
Status: CLOSED CURRENTRELEASE QA Contact: Elad <ebenahar>
Severity: medium Docs Contact:
Priority: unspecified    
Version: 3.3.0CC: aberezin, amureini, bazulay, derez, ecohen, iheim, michal.skrivanek, rbalakri, Rhev-m-bugs, scohen, sherold, tnisan, yeylon
Target Milestone: ---Flags: amureini: Triaged+
Target Release: 3.5.0   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard: storage
Fixed In Version: ovirt-3.5.0_rc1 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: Storage RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Aharon Canan 2013-11-27 13:16:58 UTC
Description of problem:
after enabling virtIO-SCSI we need to restart VM so the controller will start
no tip/comment about it to the user. 
creating a disk and trying to active it fails as no restart took place

Version-Release number of selected component (if applicable):
is24.2

How reproducible:
100

Steps to Reproduce:
1. create VM without virtio-scsi enable with one virtio disk
2. enable virtio-scsi
3. 

Actual results:
no comment

Expected results:
should ask to user to restart VM

Additional info:

Comment 2 Arthur Berezin 2014-04-29 17:17:38 UTC
Currently we should add a popup asking the user to reboot the VM if virtio-scsi was added. 
On the longer run we need to add support for for hotplug PCI, scott is this under our radar?

Comment 3 Scott Herold 2014-05-08 18:55:40 UTC
I know RHEV already supports Hot Plug of vNICs.  Is this already using PCI Hot Plug capabilities, or is it using another mechanism?  Can we use the same/similar mechanism for the SCSI controller?  If not, please create a new Virt RFE, and I will prioritize for 3.6.

Comment 4 Michal Skrivanek 2014-05-15 07:09:47 UTC
generic host device hotplug is on "virt"'s list, but this is a different thing. Something similar to hotplug nic/disk would need to be implemented for the controller. As to make it generic - well, not sure, depends on each device's limitation in libvirt/qemu, but it's surely something for "infra" to consider

Comment 5 Barak 2014-05-20 10:20:40 UTC
I assume there is a different behavior for hoptplug vNic, cpu, disk

So IIUC no restart required for vNic nor for vCpu,

So the question is why is restart required for virtIO-SCSI disk.

Allon ?

Comment 6 Allon Mureinik 2014-05-20 10:39:48 UTC
(In reply to Barak from comment #5)
> I assume there is a different behavior for hoptplug vNic, cpu, disk
> 
> So IIUC no restart required for vNic nor for vCpu,
> 
> So the question is why is restart required for virtIO-SCSI disk.
> 
> Allon ?
IIRC, it's related to the way we dynamically create a controller.
Daniel, can you answer Barak please?

Comment 7 Daniel Erez 2014-05-20 12:09:47 UTC
(In reply to Allon Mureinik from comment #6)
> (In reply to Barak from comment #5)
> > I assume there is a different behavior for hoptplug vNic, cpu, disk
> > 
> > So IIUC no restart required for vNic nor for vCpu,
> > 
> > So the question is why is restart required for virtIO-SCSI disk.
> > 
> > Allon ?
> IIRC, it's related to the way we dynamically create a controller.
> Daniel, can you answer Barak please?

A restart is needed since we create the VirtIO-SCSI controller upon VM start;
hot-plug of such controller is not supported.

Comment 8 Barak 2014-05-27 11:22:26 UTC
Looking at patch http://gerrit.ovirt.org/#/c/27582/ it looks like the same issue.

Moving to virt as this is on virt doamain and they have aleady implemented it on a different use case.

Comment 9 Michal Skrivanek 2014-05-29 08:30:34 UTC
"edit running vm" feature is in 3.5, it alerts on modification of fields which requires VM to be down. Storage needs to define an attribute as such. I talked with Daniel and he's checking which on to use (as the controller itself is not an entity ATM, but any other field can be used)

Comment 10 Elad 2014-08-11 05:44:34 UTC
Enabling/disabling virt-IO-SCSI in edit VM while the VM is running triggers a warning message in webadmin and also a graphical sign on the VM itself that says that the VM has some configuration which will be valid only after the next restart.

Verified on ovirt-3.5 rc1
 ovirt-engine-3.5.0-0.0.master.20140804172041.git23b558e.el6.noarch

Comment 11 Allon Mureinik 2015-02-16 19:13:38 UTC
RHEV-M 3.5.0 has been released, closing this bug.

Comment 12 Allon Mureinik 2015-02-16 19:13:38 UTC
RHEV-M 3.5.0 has been released, closing this bug.