Closed Bug 1211959 Opened 4 years ago Closed 3 years ago
Intermittent media tests FAIL test
_video _playback .py Test Video Playback .test _playback _starts | after Assertion Error: Timeout Exception: Timed out after 62 .8 seconds; condition: playback _started
58 bytes, text/x-review-board-request
jya: This failure spiked around October 6. Do you think something landed around then that could have regressed this?
I would have said the new PDMFactory ; but it didn't land until after that. The only commit I see that happened just at that time is bug 1211339 which could influence playback not starting
Wonder if bug 1211443 is the cause. Even though his went in one week before this bug started.
See Also: → 1211443
It'll get a bit quieter in here, now that I've hidden them on ASan for a failure rate beyond belief.
Whiteboard: [hidden on ASan]
Depends on: 1235583
This is spiking today. Any thoughts? https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1211959&startday=2016-05-02&endday=2016-06-03&tree=all
My guess is probably still Bug 1235583. Not sure why it would spike though.
I don't know this test. We have seen failures spike yesterday on webrtc tests, likely due to networking issue (infrastructure). Maybe that's related.
Maja, can you take this and try to get it fixed? Its one of our top intermittents at the moment.
My understanding is that :SingingTree is the best person to take care of it now, but I'm happy to provide guidance if needed.
Flags: needinfo?(mjzffr) → needinfo?(bvandyk)
Sure thing, I'm fairly new to these tests still, so any guidance would be appreciated. I'm seeing this fail regularly at the moment on our Windows 8 test machine, but am uncertain as to what exactly is going on thus far.
Bryce, any luck here?
Not yet. We've had a number of the VMs that the external tests were running on taken out by the MTV power outage. I recently got the credentials required to log in and get them stood back up. Hoping to get all of those VMs restored and happy in the next few days, then be able to debug some of these fails.
I've been able to recreate some (not sure if all fails are down to what I'm seeing) of these fails. All the local reproductions I've seen thus far are not that playback doesn't start, but rather that playback does start and the start event is not correctly recorded. Typically I'd see playback reach the end of a video, at which point the initial time stamp would be recorded, and then the tests would check to see if initial_time_stamp < current_time_stamp. Because the video had ended this conditional will never change and then we hit a timeout.
Comment on attachment 8778844 [details] Bug 1211959 - Tests check if video has started using played range. https://reviewboard.mozilla.org/r/69904/#review67150
Attachment #8778844 - Flags: review?(dburns) → review+
Pushed by firstname.lastname@example.org: https://hg.mozilla.org/integration/autoland/rev/d1274a312b6b Tests check if video has started using played range. r=automatedtester
Comment on attachment 8778844 [details] Bug 1211959 - Tests check if video has started using played range. Approval Request Comment [Feature/regressing bug #]: Bug 1211959 [User impact if declined]: Higher rate of failures of external media tests (https://developer.mozilla.org/en-US/docs/Mozilla/QA/external-media-tests) for Aurora branch. [Describe test coverage new/current, TreeHerder]: Reduces race conditions in external media tests. [Risks and why]: Low risk. Could break external media tests. Even in this case no direct impact would be had on end users. [String/UUID change made/needed]: None.
Comment on attachment 8778844 [details] Bug 1211959 - Tests check if video has started using played range. Automation test changes are auto-approved. These do not need relman review.
Bryce, pulsebot doesn't handle bug commenting or metadata setting for uplifts. Please don't forget to update the bug when you push an uplift in the future. https://hg.mozilla.org/releases/mozilla-aurora/rev/8bf3072e8e924fc4fbfd150daf4c2106a62bae0e
You need to log in before you can comment on or make changes to this bug.