Bug 747956

Summary: win2k3 32b guest got BSOD(0x7F) when run whql job Common_Scenario_Stress_With_IO
Product: Red Hat Enterprise Linux 5 Reporter: Qin Guan <qguan>
Component: xenpv-winAssignee: Paolo Bonzini <pbonzini>
Status: CLOSED WONTFIX QA Contact: Virtualization Bugs <virt-bugs>
Severity: high Docs Contact:
Priority: high    
Version: 5.8CC: bcao, leiwang, mrezanin, pbonzini, qwan, rhod, wshi, yuzhou, yvugenfi
Target Milestone: rcKeywords: Reopened
Target Release: ---   
Hardware: i386   
OS: Windows   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2013-06-21 14:20:27 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
error-0x2b02-whql-2k3-32b.png
none
bsod-0x7e-whql-2k3-32b.png
none
bsod-0xA-whql-2k3-32b.png
none
bsod-0xA-whql-2k3-32b.png
none
windows 2k3 BSOD screenshot
none
dump analyse none

Comment 1 Qin Guan 2011-10-21 14:03:15 UTC
Created attachment 529514 [details]
error-0x2b02-whql-2k3-32b.png

Comment 2 Qin Guan 2011-10-21 14:03:56 UTC
Created attachment 529515 [details]
bsod-0x7e-whql-2k3-32b.png

Comment 3 Paolo Bonzini 2011-10-24 08:36:06 UTC
Can you try with xenpv-win-1.3.4 and WLK1.6 to find if it is a regression in RHELSCSI or the bug is due to WLK1.6 (which could be a driver bug exposed by WLK1.6, or a bug in WLK1.6)?

Comment 4 Qin Guan 2011-10-25 04:02:23 UTC
(In reply to comment #3)
> Can you try with xenpv-win-1.3.4 and WLK1.6 to find if it is a regression in
> RHELSCSI or the bug is due to WLK1.6 (which could be a driver bug exposed by
> WLK1.6, or a bug in WLK1.6)?

Another BSOD (0xA) happen when testing xenpv-win-1.3.4 and WLK1.6. 

The steps are exactly the same as before, so does the error message (Error:0x2b02). The BSOD happened at the same phase (guest hibernate) but with different stop code 0xA (bsod-0xA-whql-2k3-32b.png).

Comment 5 Qin Guan 2011-10-25 04:03:51 UTC
Created attachment 529996 [details]
bsod-0xA-whql-2k3-32b.png

Comment 7 Qin Guan 2011-10-25 04:21:53 UTC
Created attachment 529998 [details]
bsod-0xA-whql-2k3-32b.png

Comment 8 Qin Guan 2011-10-27 13:57:43 UTC
Please ignore the error message of 0x2b02 (Error:0x2b02, Error due to lack of resources), it's gone after removing the DHCP default gateway from DHCP Server.

While the BSOD still exists.

Comment 9 Paolo Bonzini 2011-10-27 15:26:32 UTC
The two BSODs look similar, and I couldn't reproduce it with ~40 hibernations outside the common scenario stress with I/O test.  Can you give me your settings for EnumerateDevices/EnumerateDevicesOverride?  I suppose they are 1.  Can you try the test with EDO=3?  Also please attach the guest configuration.  Thanks!

Comment 18 Paolo Bonzini 2011-11-11 22:34:02 UTC
I cannot reproduce this here.

Comment 19 Paolo Bonzini 2011-11-11 22:45:34 UTC
Does hibernation work at all for you (i.e. invoked manually outside logo tests? you need to issue "powercfg /hibernate on" to see the item in the shut down menu).

Comment 20 Qin Guan 2011-11-15 12:53:47 UTC
Just lucky to run the jobs PASS with a fresh installed image on another host. 

The three jobs "Common Scenario Stress With DiskIO", "Common Scenario Stress With IO" and "Sleep Stress With IO" all PASS without BSOD. And the rest of the jobs are still running.

The hibernation works well when manually invoked, while it still reproduced on previous host even with the same fresh image. I will continue the testing to make sure the job can finally PASS, and will provide more info if there are some new findings.

Comment 24 Yuyu Zhou 2012-09-28 03:01:59 UTC
Created attachment 618377 [details]
windows 2k3 BSOD screenshot

Comment 25 Yuyu Zhou 2012-09-28 03:02:49 UTC
Created attachment 618378 [details]
dump analyse

Comment 26 Paolo Bonzini 2013-01-29 16:59:23 UTC
> The manual invoke the hibernate in Win2k3 32b guest get BSOD 100% now.
> 1. manual hibernate works well in the this bug.

I cannot understand the above, it seems like a contradiction.  Manual hibernate works well for me in w2k3 32-bit and 64-bit.

Comment 27 Yuyu Zhou 2013-02-01 03:05:55 UTC
Hello, Paolo
It was reproduced on a HCK environment but it's not available currently.
And the test pass on a new setup environment (out of SVVP).
I will try to reproduce it when running HCK testing next time. 
You could close this bug if not reproducible, and I could re-open later if I reproduced it next time.
Yuyu Zhou