New issue
Advanced search Search tips
Starred by 2 users

Issue metadata

Status: Assigned
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug
Flaky-Test: WebRtcGetUserMediaBrowserTest.ApplyConstraintsNonDevice



Sign in to add a comment
link

Issue 915135: WebRtcGetUserMediaBrowserTest.ApplyConstraintsNonDevice/* is flaky

Reported by Findit, Dec 14 Project Member

Issue description

Comment 1 by jonr...@chromium.org, Dec 14

Cc: kylec...@chromium.org jonr...@chromium.org
Looks like a limited flake on ASAN: 90% pass rate.

https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/win-asan/2544

Comment 3 by dtapu...@chromium.org, Dec 14

Components: Blink>WebRTC
Owner: kylec...@chromium.org
Status: Assigned (was: Untriaged)

Comment 4 by kylec...@chromium.org, Dec 14

Looks like the test is getting back 299 instead of 300 for the width here:

https://cs.chromium.org/chromium/src/content/test/data/media/getusermedia.html?l=796&rcl=afc14cd53bf3ef4d62107cfc5c0af26ac206faed

It's strange we didn't get this flake last week, or maybe Findit didn't have a chance to find our CL before we reverted. We could potentially disable the test on win+asan until we have a solution to see if there is any other flaky tests that are going to get discovered?

Comment 5 by jonr...@chromium.org, Dec 14

Owner: jonr...@chromium.org
With it being just win-asan I think the scope is limited enough to just disable the particular test.

I'll take a quick look locally first to see if something stands out though

Comment 6 by bugdroid1@chromium.org, Dec 14

Project Member
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/c98089352f6ab0c6d0e69341047b73028aabea4f

commit c98089352f6ab0c6d0e69341047b73028aabea4f
Author: jonross <jonross@chromium.org>
Date: Fri Dec 14 18:56:29 2018

Disable WebRtcGetUserMediaBrowserTest.ApplyConstraintsNonDevice/* on Win-ASAN

With VizDisplayCompositor being set to on by default, this test has begun
flaking on WIN-ASAN only.

As we'd like to identify other potential flakes I'm electing to temporarily
disble the test, rather than revert the feature flag. I'll begin looking at it
locally.

TBR=phoglund@chromium.org
TEST=WebRtcGetUserMediaBrowserTest.ApplyConstraintsNonDevice/*

Bug: 915135
Change-Id: I6435c556a6f6b23defa1da6976195e5d4d5bb521
Reviewed-on: https://chromium-review.googlesource.com/c/1377608
Reviewed-by: Jonathan Ross <jonross@chromium.org>
Commit-Queue: Jonathan Ross <jonross@chromium.org>
Cr-Commit-Position: refs/heads/master@{#616768}
[modify] https://crrev.com/c98089352f6ab0c6d0e69341047b73028aabea4f/content/browser/webrtc/webrtc_getusermedia_browsertest.cc

Comment 7 by hongchan@chromium.org, Dec 14

Labels: -Sheriff-Chromium
The flaky test is disabled. Removing sheriff label.

Comment 8 by jonr...@chromium.org, Dec 14

Owner: guidou@chromium.org
+guidou@ whom worked on the implementation of applyConstraints.

I can sometimes reproduce this locally, however I am not getting any LOG input out of calls to content/renderer/media/stream/apply_constraints_processor.cc
It seems that when the JS is ran that the c++ LOGs go missing. But I'm not sure what flag would get them to show.

Since this is only on win-asan, and tied to VizDisplayCompositor. I suspect that we just have multiple timing changes that are exposing a race.

Looking at WebRtcGetUserMediaBrowserTest.ApplyConstraintsNonDevice it performs a NavigateToURL, then immediately begins the javascript execution. I suspect that sometimes we are now executing that before surface synchronization has completed, and therefore before the final size has been set.

browser_test_utils::RenderFrameSubmissionObserver::WaitForMetadataChange() would prevent such a race.

guidou@ could you provide guidance on how to get the logs? I would prefer to confirm the supposition before landing a change

Comment 9 by jonr...@chromium.org, Dec 14

Apparently adding just the metadata sync is insufficient.

So we need to debug ApplyConstraintsNonDevice

Comment 11 by thestig@chromium.org, Dec 17

Cc: nisse@chromium.org guidou@chromium.org
 Issue 915695  has been merged into this issue.

Comment 12 by thestig@chromium.org, Dec 17

https://chromium-review.googlesource.com/1380964 to disable another failing test.

Comment 13 by bugdroid1@chromium.org, Dec 17

Project Member
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/8122556c9417133a20a363bd366f759dde5a4aa9

commit 8122556c9417133a20a363bd366f759dde5a4aa9
Author: Lei Zhang <thestig@chromium.org>
Date: Mon Dec 17 21:42:43 2018

Disable a WebRtcGetUserMediaBrowserTest under WinASAN.

WebRtcGetUserMediaBrowserTest.ApplyConstraintsNonDevice is flaky.

BUG=915135
TBR=guidou@chromium.org

Change-Id: I1c29c50d799866a55b4cc20f20c2aa8d6f7543d5
Reviewed-on: https://chromium-review.googlesource.com/c/1380964
Reviewed-by: Lei Zhang <thestig@chromium.org>
Commit-Queue: Lei Zhang <thestig@chromium.org>
Cr-Commit-Position: refs/heads/master@{#617240}
[modify] https://crrev.com/8122556c9417133a20a363bd366f759dde5a4aa9/content/browser/webrtc/webrtc_getusermedia_browsertest.cc

Comment 14 by bugdroid1@chromium.org, Dec 17

Project Member
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/ec2f50f3ba0f27d09b8faacae85ac5b790588cd0

commit ec2f50f3ba0f27d09b8faacae85ac5b790588cd0
Author: Jonathan Ross <jonross@chromium.org>
Date: Mon Dec 17 22:30:56 2018

Revert "Disable 
WebRtcGetUserMediaBrowserTest.ApplyConstraintsNonDevice/* on Win-ASAN"

This reverts commit c98089352f6ab0c6d0e69341047b73028aabea4f.

Reason for revert: disabled wrong test my mistake

Original change's description:
> Disable WebRtcGetUserMediaBrowserTest.ApplyConstraintsNonDevice/* on Win-ASAN
> 
> With VizDisplayCompositor being set to on by default, this test has begun
> flaking on WIN-ASAN only.
> 
> As we'd like to identify other potential flakes I'm electing to temporarily
> disble the test, rather than revert the feature flag. I'll begin looking at it
> locally.
> 
> TBR=phoglund@chromium.org
> TEST=WebRtcGetUserMediaBrowserTest.ApplyConstraintsNonDevice/*
> 
> Bug: 915135
> Change-Id: I6435c556a6f6b23defa1da6976195e5d4d5bb521
> Reviewed-on: https://chromium-review.googlesource.com/c/1377608
> Reviewed-by: Jonathan Ross <jonross@chromium.org>
> Commit-Queue: Jonathan Ross <jonross@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#616768}

TBR=phoglund@chromium.org,jonross@chromium.org

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug: 915135
Change-Id: I4848db0e27a310d99a3bea3bf6c2a0ff78b0026b
Reviewed-on: https://chromium-review.googlesource.com/c/1381131
Reviewed-by: Jonathan Ross <jonross@chromium.org>
Reviewed-by: Lei Zhang <thestig@chromium.org>
Commit-Queue: Jonathan Ross <jonross@chromium.org>
Cr-Commit-Position: refs/heads/master@{#617261}
[modify] https://crrev.com/ec2f50f3ba0f27d09b8faacae85ac5b790588cd0/content/browser/webrtc/webrtc_getusermedia_browsertest.cc

Comment 15 by thestig@chromium.org, Dec 18

Labels: Sheriff-Chromium
The main Windows builder died earlier today, and we finally brought it back. Now its running again and showing this issue without ASAN as well. Here's a sample failure: https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Win%207%20Tests%20x64%20%281%29/46109

Putting this back on the sheriff's radar.

Comment 18 by bugdroid1@chromium.org, Dec 18

Project Member
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/e56bb3b941e6e1e973dacf335e86b99d194b8f65

commit e56bb3b941e6e1e973dacf335e86b99d194b8f65
Author: Tsuyoshi Horo <horo@chromium.org>
Date: Tue Dec 18 05:14:45 2018

Disable WebRtcGetUserMediaBrowserTest.ApplyConstraintsNonDevice on Win

BUG=915135
TBR=guidou@chromium.org

Change-Id: I479ee014ff241b48e3c2b6ada9ade58b3551c866
Reviewed-on: https://chromium-review.googlesource.com/c/1381354
Reviewed-by: Tsuyoshi Horo <horo@chromium.org>
Commit-Queue: Tsuyoshi Horo <horo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#617384}
[modify] https://crrev.com/e56bb3b941e6e1e973dacf335e86b99d194b8f65/content/browser/webrtc/webrtc_getusermedia_browsertest.cc

Comment 19 by horo@chromium.org, Dec 18

Labels: -Sheriff-Chromium

Comment 20 by nisse@chromium.org, Dec 18

Regarding https://chromium-review.googlesource.com/c/chromium/src/+/1373757 (suspected culprit cl, that's been reverted). The intention is to keep the frame size behavior (including usage of webrtc's VideoAdapter class) unchanged, coping that code from the deleted class WebRtcVideoCapturerAdapter.

An off-by-one change, from 299 to 300, seems unlikely from that cl.

Comment 21 by guidou@chromium.org, Dec 18

Comment 22 by guidou@chromium.org, Dec 18

Labels: -Pri-1 Pri-2
Lowering priority since the flaky test has been disabled.

Sign in to add a comment