721166 - [cb-seamonkey-win32-0x] mochitest-plain-5: "test_videocontrols.html | Test timed out." + a leak

Status: RESOLVED WONTFIX

(SeaMonkey :: Release Engineering, defect, P2)

Description

[Serge Gautherie (:sgautherie)]

http://tinderbox.mozilla.org/showbuilds.cgi?tree=SeaMonkey&maxdate=1326867500&hours=24&legend=0&norules=1
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1326799065.1326800355.32047.gz
WINNT 5.2 comm-central-trunk debug test mochitests-5/5 on 2012/01/17 03:17:45

It looks like this was the first time this failure was reported.
But I could not say whether this is a new failure or it used to be masked somehow before (by other failures)...


http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1327474678.1327476084.18149.gz&fulltext=1
WINNT 5.2 comm-central-trunk debug test mochitests-5/5 on 2012/01/24 22:57:58
{
10223 ERROR TEST-UNEXPECTED-FAIL | /tests/toolkit/content/tests/widgets/test_videocontrols.html | Test timed out.

TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 564 bytes during test execution
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 2 instances of Mutex with size 12 bytes each (24 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of ReentrantMonitor with size 16 bytes
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of nsStringBuffer with size 8 bytes
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 2 instances of nsTArray_base with size 4 bytes each (8 bytes total)
TEST-UNEXPECTED-FAIL | automationutils.processLeakLog() | leaked 1 instance of nsThread with size 124 bytes
}

Comment 1

[Serge Gautherie (:sgautherie)]

Few greens, on cn-sea-qm-win2k3-01:
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1327262748.1327264837.22631.gz
WINNT 5.2 comm-central-trunk debug test mochitests-5/5 on 2012/01/22 12:05:48
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1327586269.1327588409.1585.gz
WINNT 5.2 comm-central-trunk debug test mochitests-5/5 on 2012/01/26 05:57:49

Lots of oranges, on cb-seamonkey-win32-0[1-3]:
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1327557356.1327558597.21698.gz
WINNT 5.2 comm-central-trunk debug test mochitests-5/5 on 2012/01/25 21:55:56

SM 2.8 and 2.7 are affected too:
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey-Aurora/1327607186.1327608458.19176.gz
WINNT 5.2 comm-aurora debug test mochitests-5/5 on 2012/01/26 11:46:26
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey-Beta/1327517324.1327518174.32661.gz
WINNT 5.2 comm-beta debug test mochitests-5/5 on 2012/01/25 10:48:44

It looks like something is wrong on our 3 cb-seamonkey-win32-0x boxes.

Comment 2

[Justin Wood (:Callek)]

Dupe of? 706894

Comment 3

[Justin Wood (:Callek)]

....Basically I really doubt those machines are misconfigured, they match MoCo machine configuration (for builders anyway), they are NOT talos machines, so might be missing some GPU Accel, etc. type of work... but either way, they are configured as correctly as we can make them.

Comment 4

[Serge Gautherie (:sgautherie)]

(In reply to Justin Wood (:Callek) from comment #2)
> Dupe of? 706894

I don't think so: SeaMonkey issue is perma, not random.


(In reply to Justin Wood (:Callek) from comment #3)
> ....Basically I really doubt those machines are misconfigured, they match
> MoCo machine configuration

I am aware of our box limitations, but this test passes on cn-sea-qm-win2k3-01 box, so there must be some difference(s) on cb-seamonkey-win32-0x boxes, must it not?
I think we should find out what it is.

Comment 5

[Serge Gautherie (:sgautherie)]

(In reply to Serge Gautherie (:sgautherie) from comment #4)
> I am aware of our box limitations, but this test passes on
> cn-sea-qm-win2k3-01 box, so there must be some difference(s) on
> cb-seamonkey-win32-0x boxes, must it not?
> I think we should find out what it is.

(Current) Bug 562626 is similar and passes on sea-win32-02.

Where are these VMs hosted?
Could it be a fallout of major Parallels (host) breakdown that happened not so long ago?

Comment 6

[Serge Gautherie (:sgautherie)]

PS:
Iiuc, all SeaMonkey VMs are/will being rebuilt on Vmware ESX.
In the meantime, to get clean test results on Windows, would it be appropriate to use the 3 cb-seamonkey-win32-0x as builders only?
And maybe (some of) the other boxes (cn-sea-qm-win2k3-01, sea-win32-02, ...) as testers only, as need be?

Comment 7

[Serge Gautherie (:sgautherie)]

(In reply to Serge Gautherie (:sgautherie) from comment #6)

> In the meantime, to get clean test results on Windows, would it be
> appropriate to use the 3 cb-seamonkey-win32-0x as builders only?

Could you do that?
Current state is counter-productive: computer and human.

> And maybe (some of) the other boxes (cn-sea-qm-win2k3-01, sea-win32-02, ...)
> as testers only, as need be?

***

Ftr,
I haven't checked (yet), but, Linux/Parallels VMs might be affected too:
cb-seamonkey-linux-01 (and cb-seamonkey-linux-02)
cb-seamonkey-linux64-01

Comment 8

[Serge Gautherie (:sgautherie)]

{
<sgautherie>	Callek: could you disable the Parallels Windows builders for mochitests (all 3 branches)?
<Callek>	sgautherie: doing so would be _hard_
<sgautherie>	Callek: oh, I hoped that would be just removing them from a scheduler list or something...
<Callek>	so figuring out how they are different/broken is helpful
<sgautherie>	Callek: Only you or ewong can investigate...
}

The other option would be to disable the failing tests, which I would rather avoid (the trouble).

Comment 9

[Serge Gautherie (:sgautherie)]

SM 2.9:

http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey-Release/1334981669.1334982646.29079.gz&fulltext=1
WINNT 5.2 comm-release debug test mochitests-5/5 on 2012/04/20 21:14:29
s: cb-seamonkey-win32-02
{
10286 INFO TEST-END | /tests/toolkit/content/tests/widgets/test_videocontrols.html | finished in 5983ms
}

http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey-Release/1335661205.1335662258.6485.gz&fulltext=1
WINNT 5.2 comm-release debug test mochitests-5/5 on 2012/04/28 18:00:05
s: cb-seamonkey-win32-01
{
10286 INFO TEST-END | /tests/toolkit/content/tests/widgets/test_videocontrols.html | finished in 6279ms
}

R.WorksForMe. (I don't know what/when fixed this.)

Comment 10

[Serge Gautherie (:sgautherie)]

http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1336466689.1336467847.19536.gz
WINNT 5.2 comm-central-trunk debug test mochitests-5/5 on 2012/05/08 01:44:49
s: cb-seamonkey-win32-03

still good.


http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1336595054.1336596573.12601.gz
WINNT 5.2 comm-central-trunk debug test mochitests-5/5 on 2012/05/09 13:24:14

Regressed again :-<

Hint:
Screenshot shows a Windows alert:
"Display Settings
 ... resolution and color depth ... very low ..."

Did the box(es) reboot? Are their default settings set too low? ...

Comment 11

[Serge Gautherie (:sgautherie)]

http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1336627050.1336627904.13291.gz
WINNT 5.2 comm-central-trunk debug test mochitests-5/5 on 2012/05/09 22:17:30

Ftr, (as expected) sea-win32-02 is not affected.

Comment 12

[Edmund Wong (:ewong)]

Marking as invalid as this system no longer exists. Decommissioned eons ago.