Closed Bug 1373116 Opened 5 years ago Closed 5 years ago

Intermittent browser/components/sessionstore/test/browser_windowStateContainer.js | application crashed [@ RunWatchdog]

Categories

(Firefox :: Session Restore, defect)

defect
Not set
critical

Tracking

()

RESOLVED INCOMPLETE

People

(Reporter: intermittent-bug-filer, Unassigned)

References

(Blocks 1 open bug)

Details

(Keywords: crash, intermittent-failure, Whiteboard: [stockwell disabled])

Crash Data

Attachments

(1 file)

Filed by: philringnalda [at] gmail.com

https://treeherder.mozilla.org/logviewer.html#?job_id=107187587&repo=autoland

https://queue.taskcluster.net/v1/task/dywoNPfeQWGJmEjL2-K4fw/runs/0/artifacts/public/logs/live_backing.log

In general, a RunWatchdog failure in CI is pretty much meaningless, since despite being blamed on a particular test that test is just the last test which ran during the browser session, and the problem RunWatchdog is proclaiming, "shutdown took longer than I wanted it to take" could have been caused by any test in the run, not just that one.

However, this is failing constantly (one set of retriggers I saw it failed 9 out of 10 runs), and in browser-chrome we start and stop the browser for every directory of tests, so this failure is actually "one of the tests in browser/components/sessionstore/test/ causes shutdown on Linux32 to take too long."
67 failures since June 14th, this is on track to be one of our top intermittents.  I see a fix in bug 1358898, I assume this will be fixed by then.
Whiteboard: [stockwell needswork]
(In reply to Joel Maher ( :jmaher) from comment #3)
> 67 failures since June 14th, this is on track to be one of our top
> intermittents.  I see a fix in bug 1358898, I assume this will be fixed by
> then.

indeed this is nearly perma failing now
this is still failing and bug bug 1358898 didn't change anything.

:mike, I see you are the triage owner for sessionrestore, can you help find someone on the sessionrestore team to look at this very frequent intermittent failure and help get a resolution (fix, reduced, disabled) by the end of  the week?
Flags: needinfo?(mdeboer)
Yoshi wrote the test and is basically the owner for it - perhaps he knows what might be going on?
But I think a good first move is the just disable this test on linux. The failure rate is just insane atm.
Flags: needinfo?(mdeboer) → needinfo?(allstars.chh)
Comment on attachment 8879908 [details]
Bug 1373116 - Disable browser_windowStateContainer.js mochitest on Linux until it is fixed.

https://reviewboard.mozilla.org/r/151312/#review156198

you might want to consider doing this on opt only:
skip-if = os == "linux" && !debug

that would get us some coverage on linux :)
Attachment #8879908 - Flags: review?(jmaher) → review+
Pushed by mdeboer@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/a541b46cb8ff
Disable browser_windowStateContainer.js mochitest on Linux until it is fixed. r=jmaher
Whiteboard: [stockwell needswork] → [stockwell disabled]
Assignee: nobody → allstars.chh
Flags: needinfo?(allstars.chh)
Depends on: 1244904
This was a shutdown crash, not necessarily associated with browser_windowStateContainer.js:

https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=108679869&lineNumber=3058

[task 2017-06-20T22:44:54.796192Z] 22:44:54     INFO - TEST-OK | browser/components/sessionstore/test/browser_windowStateContainer.js | took 8460ms
[task 2017-06-20T22:44:54.943256Z] 22:44:54     INFO - checking window state
[task 2017-06-20T22:45:06.298058Z] 22:45:06     INFO - GECKO(1577) | Completed ShutdownLeaks collections in process 1577
[task 2017-06-20T22:45:06.300415Z] 22:45:06     INFO - TEST-START | Shutdown
[task 2017-06-20T22:45:06.301639Z] 22:45:06     INFO - Browser Chrome Test Summary
[task 2017-06-20T22:45:06.304550Z] 22:45:06     INFO - Passed:  1174
[task 2017-06-20T22:45:06.306482Z] 22:45:06     INFO - Failed:  0
[task 2017-06-20T22:45:06.308723Z] 22:45:06     INFO - Todo:    0
[task 2017-06-20T22:45:06.311256Z] 22:45:06     INFO - Mode:    non-e10s
[task 2017-06-20T22:45:06.319324Z] 22:45:06     INFO - *** End BrowserChrome Test Results ***
[task 2017-06-20T22:46:11.689491Z] 22:46:11     INFO - GECKO(1577) | ExceptionHandler::GenerateDump cloned child 1666
[task 2017-06-20T22:46:11.692337Z] 22:46:11     INFO - GECKO(1577) | ExceptionHandler::SendContinueSignalToChild sent continue signal to child
[task 2017-06-20T22:46:11.694488Z] 22:46:11     INFO - GECKO(1577) | ExceptionHandler::WaitForContinueSignal waiting for continue signal...
[task 2017-06-20T22:46:12.376110Z] 22:46:12     INFO - TEST-INFO | Main app process: exit 11
[task 2017-06-20T22:46:12.378619Z] 22:46:12     INFO - Buffered messages finished
[task 2017-06-20T22:46:12.380479Z] 22:46:12    ERROR - TEST-UNEXPECTED-FAIL | browser/components/sessionstore/test/browser_windowStateContainer.js | application terminated with exit code 11


browser_windowStateContainer.js was just the last test to run in that job. Disabling the test moved the crash to bug 1244904.
Assignee: allstars.chh → nobody
Status: NEW → RESOLVED
Closed: 5 years ago
Resolution: --- → INCOMPLETE
Removing leave-open keyword from resolved bugs, per :sylvestre.
Keywords: leave-open
You need to log in before you can comment on or make changes to this bug.