Closed Bug 821420 Opened 12 years ago Closed 11 years ago

Intermittent crash during mochitests on B2G ("This usually indicates the B2G process has crashed" after a " | application timed out after 330 seconds with no output")

Categories

(Firefox OS Graveyard :: General, defect)

Product:
Firefox OS Graveyard
Component:
General
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(Not tracked)

Status:
RESOLVED FIXED

People

(Reporter: jgriffin, Unassigned)

References

Details

(Keywords: intermittent-failure) 

There is an intermittent crash which occurs during mochitests, apparently only on the b2g18 branch.  The test that can be active when the crash occurs seems totally random.

We haven't seen this crash on m-c or inbound.

Example:  https://tbpl.mozilla.org/php/getParsedLog.php?id=17908104&tree=Mozilla-B2g18

09:14:56  WARNING -  TEST-UNEXPECTED-FAIL | /tests/dom/tests/mochitest/dom-level2-core/test_documentimportnode21.html | application timed out after 330 seconds with no output
09:18:07    ERROR - Return code: 1
09:18:08    ERROR -  F/libc    (  729): Fatal signal 11 (SIGSEGV) at 0x43c00000 (code=2)
09:18:08    ERROR -  This usually indicates the B2G process has crashed
09:18:08    ERROR -  F/libc    (  770): Fatal signal 11 (SIGSEGV) at 0x00000030 (code=1)
09:18:08    ERROR -  This usually indicates the B2G process has crashed

The test harness does not detect these crashes, so it reports a timeout, but in the logcat you can see that the test aborted with these crashes, and thus caused the timeout.
Blocks: b2g-mochitests
Same as this?

https://tbpl.mozilla.org/php/getParsedLog.php?id=17942344&tree=Mozilla-Inbound

{
06:14:10 WARNING - TEST-UNEXPECTED-FAIL | /tests/dom/tests/mochitest/dom-level1-core/test_hc_nodegetnextsibling.html | application timed out after 330 seconds with no output
06:17:21 ERROR - Return code: 1
06:17:21 ERROR - F/libc ( 167): Fatal signal 7 (SIGBUS) at 0x41398be4 (code=2)
06:17:21 ERROR - This usually indicates the B2G process has crashed
06:17:21 ERROR - F/libc ( 430): Fatal signal 11 (SIGSEGV) at 0x46800000 (code=2)
06:17:21 ERROR - This usually indicates the B2G process has crashed
06:17:21 ERROR - F/libc ( 465): Fatal signal 11 (SIGSEGV) at 0x00000190 (code=1)
06:17:21 ERROR - This usually indicates the B2G process has crashed
}
Yes, apparently it isn't just on b2g18.
Summary: Intermittent crash during mochitests on b2g18 branch → Intermittent crash during mochitests on B2G
Note, comment 0 and comment 1 look like bug 819248, which has some preliminary analysis.  Suggest dup'ing.
(In reply to Chris Jones [:cjones] [:warhammer] from comment #3)
> Note, comment 0 and comment 1 look like bug 819248, which has some
> preliminary analysis.  Suggest dup'ing.

Happy to dupe if they are all that :-0
:-), even
https://tbpl.mozilla.org/php/getParsedLog.php?id=18626852&tree=Mozilla-B2g18
https://tbpl.mozilla.org/php/getParsedLog.php?id=18625217&tree=Mozilla-B2g18
https://tbpl.mozilla.org/php/getParsedLog.php?id=18624650&tree=Mozilla-B2g18
https://tbpl.mozilla.org/php/getParsedLog.php?id=18625103&tree=Mozilla-B2g18
https://tbpl.mozilla.org/php/getParsedLog.php?id=18623680&tree=Mozilla-B2g18

(Adjusting summary to catch one of the more frequent modes)
Summary: Intermittent crash during mochitests on B2G → Intermittent crash during mochitests on B2G ("This usually indicates the B2G process has crashed" seen in log, often after timeout in test_transitions_and_zoom.html, ...)
Actually, the crash in test_transitions_and_zoom.html is permaorange bustage in M6 that needs backing out.
Summary: Intermittent crash during mochitests on B2G ("This usually indicates the B2G process has crashed" seen in log, often after timeout in test_transitions_and_zoom.html, ...) → Intermittent crash during mochitests on B2G ("This usually indicates the B2G process has crashed" after a " | application timed out after 330 seconds with no output")
821420
Blocks: 818103
Summary: Intermittent crash during mochitests on B2G ("This usually indicates the B2G process has crashed" after a " | application timed out after 330 seconds with no output") → Intermittent crash during mochitests, reftests on B2G ("This usually indicates the B2G process has crashed" after a " | application timed out after 330 seconds with no output")
Depends on: 845785
https://tbpl.mozilla.org/php/getParsedLog.php?id=20146247&tree=Mozilla-B2g18

But we already have a reftest bug, bug 818103.
Summary: Intermittent crash during mochitests, reftests on B2G ("This usually indicates the B2G process has crashed" after a " | application timed out after 330 seconds with no output") → Intermittent crash during mochitests on B2G ("This usually indicates the B2G process has crashed" after a " | application timed out after 330 seconds with no output")
(In reply to Phil Ringnalda (:philor) from comment #101)
> But we already have a reftest bug, bug 818103.

Ha and I filed it too lol.

Newish profile so awesomebar being less than awesome :-)
Depends on: 843296
ahal, similar to bug 818103 - whilst this may in fact be a platform issue, the harness doesn't provide enough info here - and either way, this bug and bug 818103 are soon going to result in B2G testsuites being hidden by default again. Please can you be point for improving the harness and/or banging B2G people's heads together to take a look at this? Thank you :-)
Flags: needinfo?(ahalberstadt)
(or else find someone else happy to be point)
(In reply to Ryan VanderMeulen [:RyanVM] from comment #105)
> https://tbpl.mozilla.org/php/getParsedLog.php?id=20240194&tree=Mozilla-
> Inbound
> 
> jmaher, this one may interest you.

This looks like bug 845785, I'll disable that test as well.

(In reply to Ed Morley [:edmorley UTC+0] from comment #110)
> ahal, similar to bug 818103 - whilst this may in fact be a platform issue,
> the harness doesn't provide enough info here - and either way, this bug and
> bug 818103 are soon going to result in B2G testsuites being hidden by
> default again. Please can you be point for improving the harness and/or
> banging B2G people's heads together to take a look at this? Thank you :-)

I'm working on bug 843296 to get crash information. I'll make another post in dev.b2g, but I'm not really sure what can be done without that bug being fixed as this is hard to reproduce locally. I realize this crash is super annoying and I apologize, but *please* make sure to talk to someone like Jonas Sicking or similar before hiding these tests. I don't really know what the impact of hiding them will be with e.g our partners. My intuition is that hiding them at this stage would be a very bad idea. Anyway, I'm trying to get the situation improved and I'll do some more prodding on the platform side.
Flags: needinfo?(ahalberstadt)
Ok, thank you for chasing this up :-)
Depends on: 855717
Calling this fixed by bug 867996. We've been filing B2G mochitest timeouts individually now.
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.