Bug 767664

Summary: 150 second timeout in spice
Product: [Fedora] Fedora Reporter: Jerry James <loganjerry>
Component: spiceAssignee: Gerd Hoffmann <kraxel>
Status: CLOSED CURRENTRELEASE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: low Docs Contact:
Priority: unspecified    
Version: 16CC: alexl, cfergeau, hdegoede, jforbes, kraxel, marcandre.lureau
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2012-01-12 13:38:19 EST Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Description Jerry James 2011-12-14 11:14:34 EST
Description of problem:
I use virt-manager.  I switched my guests from VNC to spice, and am mostly very happy.  However, 4 or 5 times per day, something triggers a 150 second timeout.  During that time, all guests are unresponsive.  Also, I can't do anything on the host, because the guest that triggers the timeout still has the keyboard focus.  I have tried to pay attention to general system load, but that seems unrelated to the problem.  Yesterday afternoon I had two guest VMs running and a compile going on the host, and did not trigger the problem.  This morning I only have 1 guest VM running and both it and the host are mostly idle, yet the timeout has already triggered 3 times.

I *have* noticed that it appears to be related to redisplay.  The timeout almost always occurs when I do something that triggers a repaint of an entire window.  (XEmacs does this a lot.  I'm an XEmacs developer.  Therefore, I edit files in my guests with XEmacs...)

Also, once the 150 second timeout period passes, the guest window does not redraw itself properly.  I have to close it and open a new one.  Sadly, opening a new window causes a full repaint of the entire guest desktop, which sometimes triggers the timeout again instantly.

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

How reproducible:

Steps to Reproduce:
1. Run a guest VM with spice
2. Run a program in the guest that frequently repaints its entire window
3. Wait
Actual results:
The guest window becomes unresponsive for a 150 second time period, and the window has to be closed and reopened afterward.

Expected results:
No timeouts.

Additional info:
I see two 150 second timeout values defined in the spice sources, both in server/red_worker.c: DETACH_TIMEOUT and DISPLAY_CLIENT_TIMEOUT.
Comment 1 Jerry James 2012-01-12 13:38:19 EST
I haven't had this happen once since spice-server-0.10.0-1.fc16.x86_64 came through, so I'm going to close this bug.  Thank you!  It's quite a relief to not have that happen anymore.