The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/8c33c0a01a0f30089c4741e56c6f43703771dd90

commit 8c33c0a01a0f30089c4741e56c6f43703771dd90
Author: dgozman <dgozman@chromium.org>
Date: Tue Mar 21 22:35:13 2017

Disable WebRtcMediaRecorderTest.PeerConnection on win dbg.

BUG=703844
TBR=solenberg@chromium.org
NOTRY=true

Review-Url: https://codereview.chromium.org/2769523002
Cr-Commit-Position: refs/heads/master@{#458586}

[modify] https://crrev.com/8c33c0a01a0f30089c4741e56c6f43703771dd90/content/browser/webrtc/webrtc_media_recorder_browsertest.cc

Fredrik, could you please take ownership of fixing this test?

Not sure whether this test fails because of invalid audio parameters ("GetPreferredAudioParameters failed: 88890004"), or because of a network issue ("sendto() has failed twice returning a  transient error net::ERR_ADDRESS_INVALID").

I'm guessing the latter, because of the assert that ultimately triggers (https://cs.chromium.org/chromium/src/content/test/data/media/mediarecorder_test.html?q=mediarecorder_test.html&dr&l=444).

Therefore, assigning to kjellander@ for continued routing, and CC olka@ for input on the audio_manager_win.cc message.

FYI: Here's a link to the failing test: https://luci-logdog.appspot.com/v/?s=chromium%2Fbb%2Fchromium.win%2FWin7_Tests__dbg__1_%2F58331%2F%2B%2Frecipes%2Fsteps%2Fcontent_browsertests%2F0%2Flogs%2FWebRtcMediaRecorderTest.PeerConnection_6%2F0

Edward, are you familiar with this code or should we route it to someone like mcasas@?

No, I'm not familiar with this one.

emircan: can you have a look?

[632:3908:0321/135235.235:27466986:ERROR:audio_manager_win.cc(414)] GetPreferredAudioParameters failed: 88890004
[632:3908:0321/135235.235:27466986:ERROR:audio_manager_base.cc(264)] Invalid audio output parameters received; using fake audio path. Channels: 0, Sample Rate: 0, Bits Per Sample: 0, Frames Per Buffer: 0

The code to use AudioParameters() (all zeroes) in case GetPreferredAudioParameters failed was introduced by tommi@ in this CL: https://codereview.chromium.org/2646423005 (~2 month ago).


Looks like the bot does not support CoreAudio. Should not be related to the failure, since the test passes in the previous build (https://build.chromium.org/p/chromium.win/builders/Win7%20Tests%20%28dbg%29%281%29/builds/58330) with the same error messages.

Passing it onto mcasas@ who is more familiar with the audio recording.

mcasas@ are you currently looking at this one? Is it really a pri-1?

The potential culprit range shows nothing relevant to recorder,
and media-recording audio hasn't changed in over a year and it's
working in other bots, so it smells like it could be an infra/bot
issue.

I don't think it's a Pri-1, moreover it has no milestone attached.

Flakiness of these tests has been observed in  Issue 696838 . Addressing those flakes warrants more than P3.

Can we just try to re-enable it? All VMs in Chrome infra should be identical so if it's related to CoreAudio as mentioned in #8, it should fail in the CQ if the VMs are not supporting it.

+pschmidt to speak for the Chrome infra labs Windows machine images in relation to CoreAudio.

Re #13: #8 mentions that it is unlikely related to CoreAudio, since the test is flaky, but not failing reliably. 

+tommi basing on #8

fwiw, the error is AUDCLNT_E_DEVICE_INVALIDATED, or:

"The audio endpoint device has been unplugged, or the audio hardware or associated hardware resources have been reconfigured, disabled, removed, or otherwise made unavailable for use."

The windows vm's don't support audio.  

Peter: Thanks.

olka: Then it's pretty clear: the test should deal with this and just silently skip such tests. We can make sure to run them on our own bots in chromium.webrtc[.fyi] waterfall, where we have physical machines with sound cards.

We currently run these tests: https://cs.corp.google.com/chromium_build/scripts/slave/recipe_modules/chromium_tests/chromium_webrtc.py?rcl=379bf75cbf69331d7de1f031dd8bda4f7ccc662c&l=20 but it's easy to add more - should I add it right away?

kjellander: I'm still confused why the tests fail only *sometimes*. And I'm not really familiar with thm. If it's clear to you we can do something like here in WebRTC audio debug recording (but checking for both input and output devices): https://cs.chromium.org/chromium/src/content/browser/webrtc/webrtc_audio_debug_recordings_browsertest.cc?l=115.

Yes, let's add it to the bots with the real audio.

I'm also unable to understand why the test is not always failing if the VMs don't support audio. Maybe the fleet is not entirely homogenous? Either way I think it sounds like the right thing to do to add the check you're suggesting.

I'm adding it to the bare-metal bots here: https://chromium-review.googlesource.com/c/489986/

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/tools/build/+/d5753bebd45ad4b460d981ee7c6c362fc3f1609f

commit d5753bebd45ad4b460d981ee7c6c362fc3f1609f
Author: Henrik Kjellander <kjellander@chromium.org>
Date: Fri Apr 28 12:07:57 2017

WebRTC: Add WebRtcMediaRecorderTest to our bare-metal bots.

This test is having difficulties running on the VMs, so we'll
enable it here to ensure we cover it.

This CL also sorts the list of tests alphabetically.

BUG=703844


Change-Id: I2be976a8a4175ce237159f1e3f4b532e018d7de6
Reviewed-on: https://chromium-review.googlesource.com/489986
Reviewed-by: Edward Lesmes <ehmaldonado@chromium.org>
Reviewed-by: Henrik Kjellander <kjellander@chromium.org>
Commit-Queue: Henrik Kjellander <kjellander@chromium.org>

[modify] https://crrev.com/d5753bebd45ad4b460d981ee7c6c362fc3f1609f/scripts/slave/recipes/chromium.expected/full_chromium_webrtc_fyi_Linux_Tester.json
[modify] https://crrev.com/d5753bebd45ad4b460d981ee7c6c362fc3f1609f/scripts/slave/recipes/chromium.expected/full_chromium_webrtc_fyi_Win10_Tester.json
[modify] https://crrev.com/d5753bebd45ad4b460d981ee7c6c362fc3f1609f/scripts/slave/recipes/chromium.expected/full_chromium_webrtc_Win10_Tester.json
[modify] https://crrev.com/d5753bebd45ad4b460d981ee7c6c362fc3f1609f/scripts/slave/recipe_modules/chromium_tests/chromium_webrtc.py
[modify] https://crrev.com/d5753bebd45ad4b460d981ee7c6c362fc3f1609f/scripts/slave/recipes/chromium.expected/full_chromium_webrtc_fyi_Win7_Tester.json
[modify] https://crrev.com/d5753bebd45ad4b460d981ee7c6c362fc3f1609f/scripts/slave/recipes/chromium.expected/full_chromium_webrtc_fyi_Mac_Tester.json
[modify] https://crrev.com/d5753bebd45ad4b460d981ee7c6c362fc3f1609f/scripts/slave/recipes/chromium.expected/full_chromium_webrtc_Win7_Tester.json
[modify] https://crrev.com/d5753bebd45ad4b460d981ee7c6c362fc3f1609f/scripts/slave/recipes/chromium.expected/full_chromium_webrtc_Win8_Tester.json
[modify] https://crrev.com/d5753bebd45ad4b460d981ee7c6c362fc3f1609f/scripts/slave/recipes/chromium.expected/full_chromium_webrtc_Linux_Tester.json
[modify] https://crrev.com/d5753bebd45ad4b460d981ee7c6c362fc3f1609f/scripts/slave/recipes/chromium.expected/full_chromium_webrtc_fyi_Win8_Tester.json
[modify] https://crrev.com/d5753bebd45ad4b460d981ee7c6c362fc3f1609f/scripts/slave/recipes/chromium.expected/full_chromium_webrtc_Mac_Tester.json