Bug 1021265

Summary:	VM imported from export domain gets ac97 soundcard regardles of OS type and osinfo settings
Product:	Red Hat Enterprise Virtualization Manager	Reporter:	David Jaša <djasa>
Component:	ovirt-engine	Assignee:	Roy Golan <rgolan>
Status:	CLOSED NOTABUG	QA Contact:
Severity:	medium	Docs Contact:
Priority:	unspecified
Version:	3.3.0	CC:	acathrow, desktop-qa-list, djasa, iheim, istein, lpeer, michal.skrivanek, ofrenkel, Rhev-m-bugs, yeylon
Target Milestone:	---	Keywords:	Triaged
Target Release:	3.3.0
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:	virt
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2013-11-19 14:35:30 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:

Description David Jaša 2013-10-20 21:03:38 UTC

Description of problem:
VM imported from export domain gets ac97 soundcard regardles of OS type and osinfo settings. This bug doesn't happen to fresh VMs so it is distinct from bug 1002695.

Version-Release number of selected component (if applicable):
is19 / rhevm-backend-3.3.0-0.26.master.el6ev.noarch

How reproducible:
always

Steps to Reproduce:
1. import the template and VM with Windows 7 32/64b OS from export domain
2. start the VM, check soundcard type
3.

Actual results:
engine sets soundcard to ac97 (visible in vdsm.log)

Expected results:
engine sets soundcard to ich6

Additional info:
  * my templates were created in 3.1 setup, the VMs were created in 3.3
  * I had to import the api via API to keep them based on the template - bug 1020830

Comment 2 Roy Golan 2013-10-30 08:57:21 UTC

David if those VMs where exported with device ac97, most likely they work.
if on every import we change the device to what current configuration we have we are in a risk the VM won't load properly.

so I don't think we should consult with osinfo again on import because we expect the state of the VM as is on import.

Comment 3 David Jaša 2013-10-30 11:29:21 UTC

> David if those VMs where exported with device ac97, most likely they work.

The VMs were exported with ac97 because of bug 1002695 so this bug is caused by behaviour you prefer... (IMO it's better to always regenerate so that settings are generated always correctly)

Comment 4 Roy Golan 2013-11-06 07:10:56 UTC

(In reply to David Jaša from comment #3)

>  (IMO it's better to always regenerate so that
> settings are generated always correctly)

And regenerating means something could change (even though it could be correct) and cause unexpected problems when the VM loads. As an admin, I wouldn't expect my imported VM to load with a different device than what the VM was exported with.

Comment 5 Michal Skrivanek 2013-11-07 13:31:35 UTC

I'd also not agree with regenerating on import. Better leave VM untouched because such a change can highly impact the installed OS.
I'm suggesting WONTFIX

Comment 6 David Jaša 2013-11-19 14:35:30 UTC

As agreed in person - let's close it, the only reason for this bug is to fix consequences of other short-lived bug...