Bug 922394

Summary:	--full-screen=auto-conf doesn't turn on disabled guest monitors
Product:	Red Hat Enterprise Linux 8	Reporter:	David Jaša <djasa>
Component:	spice-vdagent-win	Assignee:	Arnon Gilboa <agilboa>
Status:	CLOSED ERRATA	QA Contact:	Desktop QE <desktop-qa-list>
Severity:	high	Docs Contact:
Priority:	high
Version:	---	CC:	acathrow, byount, cfergeau, dblechte, dyasny, italkohe, jbiddle, marcandre.lureau, sgrinber, tjamrisk
Target Milestone:	rc	Keywords:	Regression, Reopened
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	vdagent-win-0.1-17	Doc Type:	Bug Fix
Doc Text:	Previously, when using a Windows virtual machine, enabling extra monitors with arbitrary configuration could fail. Now the agent enables the monitors before updating it to an arbitrary resolution to solve this issue.	Story Points:	---
Clone Of:		Environment:
Last Closed:	2013-06-10 20:27:27 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:
Embargoed:

Description David Jaša 2013-03-16 15:38:23 UTC

Description of problem:
--full-screen=auto-conf doesn't turn on disabled guest monitors

Version-Release number of selected component (if applicable):
windows guest (7 64b in my case but should work with any)
spice-vdagent -14
qxl-win -16

How reproducible:
always

Steps to Reproduce:
1. configure windows VM to have 2+ monitor but have just first enabled in the guest
2. connect to the VM with --full-screen=auto-conf (= "Full Screen" checked in User Portal console options)
3.
  
Actual results:
only the already enabled monitor gets displayed

Expected results:
other monitors that can be enabled get enabled, too (i.e. 2 monitors of 4 monitor guest get displayed on dual-monitor client)

Additional info:

Comment 2 Christophe Fergeau 2013-03-18 11:08:17 UTC

From which client version is this a regression from? (not asking for an accurate 'last working version', just which client version this was working in when you tested it).

Comment 3 David Jaša 2013-03-18 11:37:02 UTC

IMO it's agent issue. 2.2 and 3.0 behaved correctly for sure, I don't remember 3.1 behaviour.

Comment 4 Marc-Andre Lureau 2013-03-18 18:11:07 UTC


*** This bug has been marked as a duplicate of bug 922283 ***

Comment 5 David Jaša 2013-04-19 17:02:12 UTC

I tested mingw-virt-viewer-0.5.3-24 and vdagent-win-0.1-16 and this bug still does occur while 922283 does not. In addition, RHEL spice-gtk is affected, too.

Comment 6 David Blechter 2013-04-22 20:00:06 UTC

(In reply to comment #3)
> IMO it's agent issue. 2.2 and 3.0 behaved correctly for sure, I don't
> remember 3.1 behaviour.

I see exact the same behavior on 3.1 with or without "Full Screen" being checked.
So, it is not a regression

Comment 7 Marc-Andre Lureau 2013-04-22 21:09:31 UTC

Proposed fix:

http://lists.freedesktop.org/archives/spice-devel/2013-April/013152.html

Comment 8 Arnon Gilboa 2013-04-24 12:23:21 UTC

*** Bug 953984 has been marked as a duplicate of this bug. ***

Comment 13 Bryan Yount 2013-05-22 23:01:16 UTC

On old spicec:
1. Create VM with only 1 display configured
2. Install rhev-tools-3.1-9
3. Seal the VM with sysprep utility and create a template from it
4. Create a pool of VMs from that template and configure them to have 2 displays
5. Users connect with old spicec and 2 displays appear

On virt-viewer from RHEV 3.1 (and beta from 3.2):
1. Create VM with only 1 display configured
2. Install rhev-tools-3.1-9
3. Seal the VM with sysprep utility and create a template from it
4. Create a pool of VMs from that template and configure them to have 2 displays
5. Users connect with old spicec and *only 1 display* appears. The user has to open the Windows display GUI and enable the 2nd monitor manually.

This is a big enough issue that I am formally requesting a Z stream patch to 3.1. It looks like this is solved with just a RHEV tools update, can we get that pushed out before the end of life of 3.1?

Comment 14 Bryan Yount 2013-05-22 23:09:25 UTC

Copy and paste error on Comment#13, step 5 on the virt-viewer section should say "Users connect with virt-viewer and *only 1 display* appears..."

Comment 15 Bryan Yount 2013-05-23 00:32:55 UTC

Hmm, in a recent conversation with Marc-Andre, the UserPortal might not actually be passing the auto-conf flag to virt-viewer at all. If that is the case, a different bug is needed. This is still an important issue to fix but I retract my Z stream request for now.

Comment 16 David Blechter 2013-05-23 12:26:10 UTC

(In reply to Bryan Yount from comment #15)
> Hmm, in a recent conversation with Marc-Andre, the UserPortal might not
> actually be passing the auto-conf flag to virt-viewer at all.

the equivalent option in user portal exist and it calls "Full Screen".

can customer test it with the latest guest tool from rhevm 3.2?


> If that is the
> case, a different bug is needed. This is still an important issue to fix but
> I retract my Z stream request for now.

Comment 17 Bryan Yount 2013-05-25 01:37:52 UTC

(In reply to David Blechter from comment #16)
> can customer test it with the latest guest tool from rhevm 3.2?

I will have the customer test with the 3.2 beta guest tools next week. In the meantime, I have opened bug#967154 to deal with virt-viewer not automatically enabling the 2nd monitor.

Comment 20 errata-xmlrpc 2013-06-10 20:27:27 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.

http://rhn.redhat.com/errata/RHSA-2013-0924.html