Closed Bug 948895 Opened 7 years ago Closed 7 years ago

Intermittent raise ScriptTimeoutException(message=message, status=status, stacktrace=stacktrace) errors.ScriptTimeoutException: timed out

Categories

(Testing :: Marionette, defect)

x86
Linux
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED INCOMPLETE

People

(Reporter: cbook, Unassigned)

References

()

Details

Attachments

(1 file)

b2g_emulator_vm b2g-inbound opt test marionette-webapi on 2013-12-10 23:04:23 PST for push 67bcb37ba888

slave: tst-linux64-ec2-111

https://tbpl.mozilla.org/php/getParsedLog.php?id=31790766&tree=B2g-Inbound



    raise ScriptTimeoutException(message=message, status=status, stacktrace=stacktrace)
errors.ScriptTimeoutException: timed out
The instances in comment 40 to comment 51 with test_mobile_data_connection.js is the failure message were actually bustage, fixed by backout:
https://hg.mozilla.org/mozilla-central/rev/654854b387dd

Wes/Tomcat, please can we be extra careful to double check failures that seem to have suddenly spiked, to check that they are not now perma-orange :-)
We're hitting this really frequently lately. Can we get someone to look at it please?

11:12:55     INFO -  starting httpd
11:12:55     INFO -  running webserver on http://10.134.57.180:44700/
11:15:50     INFO -  waiting for system-message-listener-ready...
11:16:20     INFO -  done
11:16:38     INFO -  waiting for homescreen...
11:17:40     INFO -  Traceback (most recent call last):
11:17:40     INFO -    File "/builds/slave/test/build/tests/marionette/marionette/runtests.py", line 33, in <module>
11:17:40     INFO -      cli()
11:17:40     INFO -    File "/builds/slave/test/build/tests/marionette/marionette/runtests.py", line 28, in cli
11:17:40     INFO -      runner = startTestRunner(runner_class, options, tests)
11:17:40     INFO -    File "/builds/slave/test/build/tests/marionette/marionette/runtests.py", line 19, in startTestRunner
11:17:40     INFO -      runner.run_tests(tests)
11:17:40     INFO -    File "/builds/slave/test/build/tests/marionette/marionette/runner/base.py", line 725, in run_tests
11:17:40     INFO -      self.run_test(test)
11:17:40     INFO -    File "/builds/slave/test/build/tests/marionette/marionette/runner/base.py", line 770, in run_test
11:17:40     INFO -      self.marionette.emulator.wait_for_homescreen(self.marionette)
11:17:40     INFO -    File "/builds/slave/test/build/tests/marionette/marionette/emulator.py", line 363, in wait_for_homescreen
11:17:40     INFO -      });""", script_timeout=60000)
11:17:40     INFO -    File "/builds/slave/test/build/tests/marionette/marionette/marionette.py", line 1149, in execute_async_script
11:17:40     INFO -      filename=os.path.basename(frame[0]))
11:17:40     INFO -    File "/builds/slave/test/build/tests/marionette/marionette/marionette.py", line 605, in _send_message
11:17:40     INFO -      self._handle_error(response)
11:17:40     INFO -    File "/builds/slave/test/build/tests/marionette/marionette/marionette.py", line 654, in _handle_error
11:17:40    ERROR -      raise ScriptTimeoutException(message=message, status=status, stacktrace=stacktrace)
11:17:40    ERROR -  errors.ScriptTimeoutException: ScriptTimeoutException: timed out
11:17:42    ERROR - Return code: 1
Flags: needinfo?(overholt)
This is the script that's timing out.  We might not be getting the mozbrowserloadend event, or else it's taking longer than it used to.  (We could try increasing the timeout.)
(In reply to Jonathan Griffin (:jgriffin) from comment #109)
> This is the script that's timing out.  We might not be getting the
> mozbrowserloadend event, or else it's taking longer than it used to.  (We
> could try increasing the timeout.)

Oops, the link to the script is here:

http://mxr.mozilla.org/mozilla-central/source/testing/marionette/client/marionette/emulator.py#358
Let's see if increasing the timeout helps this problem, and add some logging in case it doesn't.
Attachment #8367694 - Flags: review?(ahalberstadt)
Assignee: nobody → jgriffin
Comment on attachment 8367694 [details] [diff] [review]
Increase wait-for-homescreen timeout,

Review of attachment 8367694 [details] [diff] [review]:
-----------------------------------------------------------------

Looks good. I had to increase this very timeout to get Mnw working on debug emulators, so I wouldn't be surprised if a minute is too short :(
Attachment #8367694 - Flags: review?(ahalberstadt) → review+
https://hg.mozilla.org/integration/mozilla-inbound/rev/70d91614a191
Whiteboard: [leave open]
Target Milestone: --- → mozilla29
I'm pretty optimistic my patch fixed this, as there haven't been any occurrences on inbound since it landed.
Assignee: jgriffin → nobody
(In reply to Jonathan Griffin (:jgriffin) from comment #168)
> I'm pretty optimistic my patch fixed this, as there haven't been any
> occurrences on inbound since it landed.

There are some comments with inbound jobs mentioned after comment #168 ... does that mean your patch didn't fix this?
Flags: needinfo?(overholt)
(In reply to Andrew Overholt [:overholt] from comment #195)
> (In reply to Jonathan Griffin (:jgriffin) from comment #168)
> > I'm pretty optimistic my patch fixed this, as there haven't been any
> > occurrences on inbound since it landed.
> 
> There are some comments with inbound jobs mentioned after comment #168 ...
> does that mean your patch didn't fix this?

No, it means there are several causes of this problem, and my patch fixed the most general cause (a failure during startup), but didn't address the specific problems that can be related to individual tests, which still occur.

Ideally, these more specific failures would be filed as distinct bugs, but I think it can be hard for sheriffs to distinguish them.
Filed bug 967318 to help us report the errors better.