Closed Bug 563561 Opened 15 years ago Closed 15 years ago

talos fails to capture a mini dump for busted ts

Categories

(Testing :: Mochitest, defect)

defect
Not set
major

Tracking

(Not tracked)

RESOLVED DUPLICATE of bug 486123

People

(Reporter: gal, Unassigned)

Details

Brief log: NOISE: __startTimestamp1272935708189__endTimestamp NOISE: NOISE: __FAILbrowser non-zero return code (35584)__FAIL Failed ts: Stopped Mon, 03 May 2010 18:15:13 FAIL: Busted: ts FAIL: browser non-zero return code (35584) Completed test ts: Stopped Mon, 03 May 2010 18:15:13 Running test tdhtml: Started Mon, 03 May 2010 18:15:13 Screen width/height:1280/1024 colorDepth:24 Browser inner width/height: 1024/618 Browser outer width/height: 1024/768 NOISE: Cycle 1: loaded http://localhost/page_load_test/dhtml/colorfade.html (next: http://localhost/page_load_test/dhtml/diagball.html) full: NOISE: __startTimestamp1272935676091__endTimestamp NOISE: __startSecondTimestamp1272935676248__endSecondTimestamp NOISE: __start_report611__end_report NOISE: NOISE: __startTimestamp1272935692134__endTimestamp NOISE: __startSecondTimestamp1272935692289__endSecondTimestamp Segmentation fault (core dumped) NOISE: __start_report617__end_report NOISE: NOISE: __startTimestamp1272935708189__endTimestamp NOISE: NOISE: __FAILbrowser non-zero return code (35584)__FAIL NOISE: __start_report617__end_report NOISE: NOISE: __startTimestamp1272935708189__endTimestamp NOISE: NOISE: __FAILbrowser non-zero return code (35584)__FAIL Failed ts: Stopped Mon, 03 May 2010 18:15:13 FAIL: Busted: ts FAIL: browser non-zero return code (35584) Completed test ts: Stopped Mon, 03 May 2010 18:15:13 The segfault can be seen in the full log, but no stack dump. I can't reproduce the crash offline, so having this dump would be very useful. This is the build that triggered this: http://hg.mozilla.org/try/rev/4c2a1dd33ac6 I am probably filing this under the wrong component. Please point me the right way if that's the case.
Severity: normal → major
What platform was this on?
The paste above is from macosx but all platforms expose the same behavior I think.
Links to the logs in question would be really useful here.
Re-run the build if needed, thats why I linked it.
Was this a try server only run? Did it ever make it to the production build side?
Tryserver only obviously since it crashes several columns ...
I will search the try server column logs. I can probably find the build logs.
I could only find that log for a linux failure resulting in a core dump, there was a timeout on vista that didn't result in a core dump and a few unit test fails as well.
Alright, I found a log. I am not sure this is exactly the same build, but take a look: http://tinderbox.mozilla.org/showlog.cgi?tree=MozillaTry&errorparser=unix&logfile=1272935019.46.1272936247.7048.gz&buildtime=1272935019.46&buildname=Linux%20try%20talos&fulltext= If you search for "busted" you can see that in the middle of the "NOISE:" block the browser dies and the slave prints "Busted ts". A little bit above that you can see the seg fault, but the mini dump isn't picked up. I can also re-run the build for you if that helps (I have a local copy, easier for me to push from here). Let me know and thanks for looking into this.
Maybe we are looking at the same log? I got this repeatedly. I will push the changeset again and we can see what happens.
Should have remembered sooner that we have trouble with symbols/crash reporting on try instead of getting your hopes up here. One thing that you can do is make a request to releng for access to one of the test boxes so that you can test directly on one of the machines - then you could track what is actually happening.
Status: NEW → RESOLVED
Closed: 15 years ago
Resolution: --- → DUPLICATE
Ok, cool, I guess we found the cause. Thanks.
Component: BrowserTest → Mochitest
You need to log in before you can comment on or make changes to this bug.