First failure seen in https://chromium-review.googlesource.com/c/chromium/src/+/1083883/3 and it looks to be related.

khushalsagar@ Can you take a look if it's related? Thanks!

Going to disable the test.

chongz@, that change hasn't landed yet. Did you mean to link to a different CL?

Re #c2: Hmmm that's wired because the flakes dashboard linked your patch as the first occurrence. Maybe the flakiness happened before your patch but you happened to run the win_optional_gpu_tests_rel trybot.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/adc0b3bcc53b2fbfddd406df5cd4c2864555134f

commit adc0b3bcc53b2fbfddd406df5cd4c2864555134f
Author: Chong Zhang <chongz@chromium.org>
Date: Mon Jun 04 20:05:11 2018

Disable DecodeVariations/VideoDecodeAcceleratorParamTest.TestSimpleDecode on Win

TBR=khushalsagar@chromium.org

Bug: 849368
Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel
Change-Id: I6cd57fd06d9de65f11b249f00131dc45ecf6b572
Reviewed-on: https://chromium-review.googlesource.com/1085906
Commit-Queue: Chong Zhang <chongz@chromium.org>
Reviewed-by: Chong Zhang <chongz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#564203}
[modify] https://crrev.com/adc0b3bcc53b2fbfddd406df5cd4c2864555134f/media/gpu/video_decode_accelerator_unittest.cc

I'm quite certain its not that change. Whatever its doing is behind a flag anyway.

chongz@, could you triage this again please, or assign to the test owner?

Cannot find an obvious culprit.

hiroh@ Can you take a look if it's related to your recent changes? Or re-assign to test owners if not?

Thanks!

Because this flakiness start happening last few days, I think this is not related to my change.
From the log, it looks like creating surface in win VDA failed. In addition, because there has been no failure on other platforms, it seems this regression is due to windows vda or graphics change.

liberato@, do you have any idea about the following error?
VDA unittest on windows fails due to this error and this failure happens in last few days.

[4460:6960:0604/041800.039:694046:ERROR:dxva_picture_buffer_win.cc(221)] Failed to create surface
[4460:6960:0604/041800.039:694046:ERROR:dxva_picture_buffer_win.cc(24)] Error in dxva_picture_buffer_win.cc on line 221
[4460:6960:0604/041800.039:694046:ERROR:mf_helpers.cc(14)] Error in dxva_video_decode_accelerator_win.cc on line 1216

 Issue 849605  has been merged into this issue.

MidStreamReset/VideoDecodeAcceleratorParamTest.TestSimpleDecode/0 is similarly flaky.

Note that there is correlation between DecodeVariations/VideoDecodeAcceleratorParamTest.TestSimpleDecode/0 flaking in a try run and DecodeVariations/VideoDecodeAcceleratorParamTest.TestSimpleDecode/0

https://chromium-try-flakes.appspot.com/all_flake_occurrences?key=ahVzfmNocm9taXVtLXRyeS1mbGFrZXNyTAsSBUZsYWtlIkFNaWRTdHJlYW1SZXNldC9WaWRlb0RlY29kZUFjY2VsZXJhdG9yUGFyYW1UZXN0LlRlc3RTaW1wbGVEZWNvZGUvMAw
vs
https://chromium-try-flakes.appspot.com/all_flake_occurrences?key=ahVzfmNocm9taXVtLXRyeS1mbGFrZXNyTgsSBUZsYWtlIkNEZWNvZGVWYXJpYXRpb25zL1ZpZGVvRGVjb2RlQWNjZWxlcmF0b3JQYXJhbVRlc3QuVGVzdFNpbXBsZURlY29kZS8wDA

 Issue 849722  has been merged into this issue.

 Issue 849701  has been merged into this issue.

 Issue 849700  has been merged into this issue.

i'm unaware of any changes to the media pipeline that would affect this.  actually, there haven't been any changes in media/gpu/windows/ since may 30.

+geofflang: has there been any changes in angle that might affect this?

Looking at the logs of win_optional_gpu_tests_rel: https://ci.chromium.org/p/chromium/builders/luci.chromium.try/win_optional_gpu_tests_rel?limit=200

There is a pretty clear cut off when this test failed on every run starting late on saturday night.  I don't see any ANGLE CLs landing around that time.  

For what it's worth, this is only on ANGLE's OpenGL backend on Windows on Intel, can we tune the test skip to just that test?

+zmo for any potential changes in surfaces/direct composition on Windows recently.

Sunny and Maggie landed a few CLs, but I can't imagine any of these will cause flakiness ...

CC them just in case

Any update?
It is not good situation that VDA unittest is out of Chrome CQ.

here's the interesting bit from win_optional_gpu_tests_rel:

(lots of failures of the VDA test)
Failed	2889	Gerrit CL 1083865 (ps#2) by skia-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com
Success	2888	Gerrit CL 1076324 (ps#10) by sadrul@chromium.org
Failed	2887	Gerrit CL 1075859 (ps#9) by sadrul@chromium.org
Failed  2886	Gerrit CL 1075859 (ps#9) by sadrul@chromium.org
Success	2885	Gerrit CL 1066304 (ps#29) by sdy@chromium.org
(no failures earlier of the VDA test)

the changes between sadrul's 1075859 ps 8 (not shown, run days earlier) and ps9 (earliest failure) don't look to be related.  so, it's likely that the failure was some other earlier commit.

i'll see what happened between there and run 2885.