New issue
Advanced search Search tips

Issue 858920 link

Starred by 2 users

Issue metadata

Status: WontFix
Owner: ----
Closed: Jul 2
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Windows
Pri: 2
Type: Bug

Blocking:
issue 786230
issue 773617
issue 859366



Sign in to add a comment

webkit_layout_tests failing reliably on vm143-m4

Project Member Reported by kbr@chromium.org, Jun 29 2018

Issue description

This VM is failing webkit_layout_tests all the time:
https://chromium-swarm.appspot.com/bot?id=vm143-m4&sort_stats=total%3Adesc

Please see this tryjob for example:
https://ci.chromium.org/p/chromium/builders/luci.chromium.try/win7_chromium_rel_ng/27572

and this shard:
https://chromium-swarm.appspot.com/task?id=3e61e4c7c0afa010&refresh=10&show_raw=1

I just clicked the "Shutdown gracefully" button in the Swarming UI, so this is only P2. Here's the log excerpt. May need to be reimaged.

18:15:23.608 3376 Checking system dependencies ...
18:15:24.107 3376 "e:\b\swarm_slave\w\ir\out\Release\content_shell.exe --check-layout-test-sys-deps" took 0.50s
18:15:24.107 3376 System dependencies check failed.
18:15:24.107 3376 To override, invoke with --nocheck-sys-deps
18:15:24.107 3376 
18:15:24.107 3376 
DevTools listening on ws://127.0.0.1:52775/devtools/browser/66a3e2a8-f617-4589-bab0-b6830cf40bbc
[5576:9248:0628/181523.842:ERROR:gpu_process_host.cc(586)] BrowserMainRunner::ExitedMainMessageLoop()
[5576:9248:0628/181523.857:WARNING:discardable_shared_memory_manager.cc(431)] Some MojoDiscardableSharedMemoryManagerImpls are still alive. They will be leaked.
[5576:8500:0628/181523.873:WARNING:scoped_temp_dir.cc(23)] Could not delete temp dir in dtor.
[5576:4468:0628/181523.873:FATAL:at_exit.cc(63)] Check failed: !g_top_manager->processing_callbacks_. 
Backtrace:
	base::debug::StackTrace::StackTrace [0x028B05B0+32]
	base::debug::StackTrace::StackTrace [0x028AFD6D+13]
	logging::LogMessage::~LogMessage [0x028C8523+83]
	base::AtExitManager::RegisterTask [0x028A970A+122]
	base::AtExitManager::RegisterCallback [0x028A9678+184]
	base::internal::CompleteLazyInstance [0x028C7AEF+31]
	disk_cache::File::Init [0x03C7198F+491]
	disk_cache::File::Init [0x03C7186D+201]
	disk_cache::MappedFile::Init [0x03C7243E+102]
	disk_cache::BlockFiles::OpenBlockFile [0x03C6AEB4+202]
	disk_cache::BlockFiles::Init [0x03C6AC50+164]
	disk_cache::BackendImpl::SyncInit [0x031961A6+430]
	disk_cache::BackendIO::ExecuteBackendOperation [0x03C68D96+58]
	base::debug::TaskAnnotator::RunTask [0x028B0DD2+306]
	base::internal::IncomingTaskQueue::RunTask [0x030D9EAC+108]
	base::MessageLoop::RunTask [0x028D1637+519]
	base::MessageLoop::DeferOrRunPendingTask [0x028D198D+157]
	base::MessageLoop::DoWork [0x028D1BBA+506]
	base::MessagePumpForIO::DoRunLoop [0x028D3E35+133]
	base::MessagePumpWin::Run [0x028D316E+110]
	base::MessageLoop::Run [0x028D1154+116]
	base::RunLoop::Run [0x028EC9FC+204]
	base::Thread::Run [0x02903224+164]
	base::Thread::ThreadMain [0x029034AD+621]
	base::PlatformThread::SetCurrentThreadPriority [0x02901675+533]
	BaseThreadInitThunk [0x7596338A+18]
	RtlInitializeExceptionChain [0x76F69F72+99]
	RtlInitializeExceptionChain [0x76F69F45+54]

18:15:24.107 3376 
18:15:24.107 3376 For complete build requirements, please see:
18:15:24.122 3376 https://chromium.googlesource.com/chromium/src/+/master/docs/windows_build_instructions.md
18:15:24.122 3376 
18:15:24.122 3376 Testing completed. Exit status: 252

 
Is there any way of getting multi-thread bt? This check failure basically looks like it's both starting up and exiting simultaneously (that, or something scribbled over AtExitTaskManager's memory)

#0: this bot passed multiple shards of webkit_layout_tests yesterday:

https://chromium-swarm.appspot.com/task?id=3e610890bbe4cb10
https://chromium-swarm.appspot.com/task?id=3e612821550e7710
https://chromium-swarm.appspot.com/task?id=3e606bbccc6a0e10

Are there particular reasons to suspect this is machine-specific vs just layout test flakiness?

Comment 3 by kbr@chromium.org, Jun 29 2018

I may have misdiagnosed based on seeing these three nearly back-to-back webkit_layout_tests runs:

https://chromium-swarm.appspot.com/task?id=3e611a2411f56f10&refresh=10&show_raw=1
https://chromium-swarm.appspot.com/task?id=3e60f62209289210&refresh=10&show_raw=1
https://chromium-swarm.appspot.com/task?id=3e60e41e1bd10810&refresh=10&show_raw=1

in the attached view.

Thought they were all the same problem but didn't dig deeply enough.

Not sure whether we should file another bug about the AtExitManager failure, which seems to happen if the harness shuts down immediately upon startup.

Screen Shot 2018-06-29 at 2.26.25 PM.png
525 KB View Download
Blockedon: 859366
Blockedon: -859366
Blocking: 859366
Sorry about misdiagnosing this. Please revive vm143-m4 and close this as WontFix; will use  Issue 859366  to track the bug in the system dependencies check. Thanks.

Status: WontFix (was: Untriaged)
just powercycled vm143-m4. 

Sign in to add a comment