First failure: https://uberchromegw.corp.google.com/i/chromium.chromiumos/builders/Linux%20ChromiumOS%20Tests%20%281%29/builds/28021

WallpaperManagerBrowserTest.SmallChildWallpaper (run #1):
[ RUN      ] WallpaperManagerBrowserTest.SmallChildWallpaper
[19056:19056:1018/123913:WARNING:audio_manager.cc(317)] Multiple instances of AudioManager detected
[19056:19056:1018/123913:WARNING:audio_manager.cc(278)] Multiple instances of AudioManager detected
Xlib:  extension "RANDR" missing on display ":9".
[19056:19073:1018/123913:WARNING:freezer_cgroup_process_manager.cc(59)] Cgroup freezer does not exist or is not writable. Unable to freeze renderer processes.
[19056:19056:1018/123913:FATAL:feature_session_type.cc(28)] Check failed: g_current_session_type == kDefaultSessionType || session_type == g_current_session_type.
#0 0x000002515c3e base::debug::StackTrace::StackTrace()
#1 0x0000025278ba logging::LogMessage::~LogMessage()
#2 0x000003256212 extensions::SetCurrentFeatureSessionType()
#3 0x0000018c4c26 chromeos::ChromeUserManagerImpl::UpdateOwnership()
#4 0x0000013cb609 content::NotificationServiceImpl::Notify()
#5 0x000001967133 chromeos::DeviceSettingsService::HandleCompletedOperation()
#6 0x00000196b234 chromeos::SessionManagerOperation::StorePublicKey()
#7 0x00000196bbfd _ZN4base8internal7InvokerINS0_9BindStateIMN8chromeos23SessionManagerOperationEFvRKNS_8CallbackIFvvELNS0_8CopyModeE1ELNS0_10RepeatModeE1EEE13scoped_refptrIN9ownership9PublicKeyEEEJNS_7WeakPtrIS4_EES9_EEEFvSF_EE3RunEPNS0_13BindStateBaseEOSF_
#8 0x00000196bce4 base::internal::ReplyAdapter<>()
#9 0x000002564803 base::(anonymous namespace)::PostTaskAndReplyRelay::RunReplyAndSelfDestruct()
#10 0x00000258eeb3 base::debug::TaskAnnotator::RunTask()
#11 0x00000252d68c base::MessageLoop::RunTask()
#12 0x00000252d928 base::MessageLoop::DeferOrRunPendingTask()
#13 0x00000252dc0b base::MessageLoop::DoWork()
#14 0x00000252f0ba base::(anonymous namespace)::WorkSourceDispatch()
#15 0x7fc823b88d13 g_main_context_dispatch
#16 0x7fc823b89060 <unknown>
#17 0x7fc823b89124 g_main_context_iteration
#18 0x00000252ef76 base::MessagePumpGlib::Run()
#19 0x00000252d438 base::MessageLoop::RunHandler()
#20 0x000002547b5e base::RunLoop::Run()
#21 0x0000006a2370 chromeos::wallpaper_manager_test_utils::WaitAsyncWallpaperLoadFinished()
#22 0x00000069fb95 chromeos::WallpaperManagerBrowserTest_SmallChildWallpaper_Test::RunTestOnMainThread()
#23 0x000001b7edb7 InProcessBrowserTest::RunTestOnMainThreadLoop()
#24 0x000001ed4aa4 content::BrowserTestBase::ProxyRunTestOnMainThreadLoop()
#25 0x000001b9183c ChromeBrowserMainParts::PreMainMessageLoopRunImpl()
#26 0x000001b909b1 ChromeBrowserMainParts::PreMainMessageLoopRun()
#27 0x0000017e7ba0 chromeos::ChromeBrowserMainPartsChromeos::PreMainMessageLoopRun()
#28 0x000001235515 content::BrowserMainLoop::PreMainMessageLoopRun()
#29 0x0000014dc3f6 content::StartupTaskRunner::RunAllTasksNow()
#30 0x000001233b31 content::BrowserMainLoop::CreateStartupTasks()
#31 0x00000123819d content::BrowserMainRunnerImpl::Initialize()
#32 0x000001231418 content::BrowserMain()
#33 0x000001b6fd94 content::ContentMainRunnerImpl::Run()
#34 0x000001b6e9c0 content::ContentMain()
#35 0x000001ed448d content::BrowserTestBase::SetUp()
#36 0x000001b7dcec InProcessBrowserTest::SetUp()
#37 0x000002384ba8 testing::Test::Run()
#38 0x0000023858e3 testing::TestInfo::Run()
#39 0x000002385d77 testing::TestCase::Run()
#40 0x00000238cf07 testing::internal::UnitTestImpl::RunAllTests()
#41 0x00000238cb5a testing::UnitTest::Run()
#42 0x000002504461 base::TestSuite::Run()
#43 0x0000005ad387 InteractiveUITestSuiteRunner::RunTestSuite()
#44 0x000001ed89af content::LaunchTests()
#45 0x0000005ad318 main
#46 0x7fc81f92e7ed __libc_start_main
#47 0x00000049c68d <unknown>

xiyuan - maybe related to session state refactoring?

Same failure occurs with WallpaperManagerBrowserTest.LargeChildWallpaper starting from the next run.

I see the same failure locally, on master at r426043, but it's flaky.

WallpaperManagerBrowserTest.LargeChildWallpaper is affected as well.

https://codereview.chromium.org/2416253004 touched session states, but it landed several hours before the flake started. I tried reverting it locally and still get the failures.

I think huangs was running a test https://codereview.chromium.org/2431433003/
to see if this was caused by https://chromiumcodereview.appspot.com/2421703002/
and https://codereview.chromium.org/2427863002

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/8c3a6c6097f2b58f77efbdad40e9ee46167e0fb7

commit 8c3a6c6097f2b58f77efbdad40e9ee46167e0fb7
Author: Steven Holte <holte@google.com>
Date: Tue Oct 18 23:00:15 2016

Disable SmallChildWallpaper test

BUG= 657180 
TBR=xdai

Review URL: https://codereview.chromium.org/2426093003 .

Cr-Commit-Position: refs/heads/master@{#426076}

[modify] https://crrev.com/8c3a6c6097f2b58f77efbdad40e9ee46167e0fb7/chrome/browser/chromeos/login/users/wallpaper/wallpaper_manager_browsertest.cc

Since the third patch in https://codereview.chromium.org/2431433003/ passed (at least on ozone bot) I believe that suggest https://chromiumcodereview.appspot.com/2421703002/ is NOT the culprit.

Actually it looks like http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_chromeos_rel_ng/builds/298026 did fail, so I will revert those CLs.

It seems like at least https://codereview.chromium.org/2427863002 might depend on those two CLs.  Since git cl upload is not working right now, I don't think I want to try unravelling all of this, since I'm not really sure what these patches are doing.  I'm leaving things as-is, with the tests disabled for now.

I think we should revert everything because:
- The message
WARNING:freezer_cgroup_process_manager.cc(59)] Cgroup freezer does not exist or is not writable. Unable to freeze renderer processes.
also appears for current failures in PushMessagingBrowserTest.SubscribeWorker

https://build.chromium.org/p/chromium.chromiumos/builders/Linux%20ChromiumOS%20Tests%20%281%29/builds/28063

- Experiment https://codereview.chromium.org/2431433003/ shows that adding 
http://crrev.com/2427863002 and http://crrev.com/2421703002 to baseline 
produces the bugs.

- This time of mysterious action-at-distance bug are hard to debug. Better to nip this now, than causing pain in the future.

I'll create a separate CL to enable tests and revert the affected CLs all at once; specifically:

http://crrev.com/2427863002
http://crrev.com/2421703002
http://crrev.com/2427863002

 Issue 657374  has been merged into this issue.

 Issue 657367  has been merged into this issue.

CL: http://crrev.com/2438533002

Just verified that the WallpaperManagerBrowserTest.SmallChildWallpaper would flake if reintroduced (See PS#1, noting that test appears successful due to retry).

Full reverts will be too messy, so I'm just going to remove selected changes of http://crrev.com/2421703002, and just disable the associated new tests:

V4LocalDatabaseManagerTest.PerformFullHashCheckCalledAsync
V4LocalDatabaseManagerTest.UsingWeakPtrDropsCallback

The CLs that make the test flaky actually expose a real problem with the test. The test login a user then mark the user as child. The user type change results in two different session type for the extension features. We are lucky that the test runs fine before.

Let me figure out a way to fix the problem.

Okay. Our experiments turned up inconclusive. I'll leave this to you then. Thanks!

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/4b2c7bf96730778554c4978e8103545965bdec22

commit 4b2c7bf96730778554c4978e8103545965bdec22
Author: xiyuan <xiyuan@chromium.org>
Date: Wed Oct 19 21:11:27 2016

cros: Use "regular" extension session type for child users

Test failures in  http://crbug.com/657180  reveal a problem of
what session type to use for users changing the child status
from the extension feature's perspective. This is a valid case
as the child status of a user could change when the user's
Gaia account is updated. Since child users are allowed to use
extensions as regular users, use "regular" extension session
type for them too.

Also re-enable the tests.

BUG= 657180 

Review-Url: https://chromiumcodereview.appspot.com/2433873005
Cr-Commit-Position: refs/heads/master@{#426284}

[modify] https://crrev.com/4b2c7bf96730778554c4978e8103545965bdec22/chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc
[modify] https://crrev.com/4b2c7bf96730778554c4978e8103545965bdec22/chrome/browser/chromeos/login/users/wallpaper/wallpaper_manager_browsertest.cc

Picked up in build 28092 (
https://build.chromium.org/p/chromium.chromiumos/builders/Linux%20ChromiumOS%20Tests%20%281%29/builds/28092)

and is green for the next 47 builds (the last one is build 28139)