New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 741936 link

Starred by 3 users

Issue metadata

Status: Duplicate
Merged: issue 580537
Owner:
Closed: Mar 2018
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug



Sign in to add a comment

BlockingLoginTest.LoginBlocksForUser is timing out frequently on Linux ChromiumOS Tests (dbg)

Project Member Reported by glevin@chromium.org, Jul 12 2017

Issue description

Not sure when this started happening, because it usually passed on retest, and didn't fail the builder.  But it's become more frequent, and has caused 3 of the last 6 runs to fail.  See

https://build.chromium.org/p/chromium.chromiumos/builders/Linux%20ChromiumOS%20Tests%20%28dbg%29%281%29/builds/28106

Also, 28109, 28110.  Further, if you open the browser_tests:stdout log for any recent run, you'll likely see at least one time out of this test.


[ RUN      ] BlockingLoginTestInstance/BlockingLoginTest.LoginBlocksForUser/7
HTTPS server started on 127.0.0.1:60517...
sending server_data: {"host": "127.0.0.1", "port": 60517} (36 bytes)
Xlib:  extension "RANDR" missing on display ":99".
[30086:30229:0711/103504.621928:WARNING:freezer_cgroup_process_manager.cc(62)] Cgroup freezer does not exist or is not writable. Unable to freeze renderer processes.
[30244:30244:0711/103505.483927:ERROR:gl_surface.cc(198)] Not implemented reached in virtual void gl::GLSurface::SetRelyOnImplicitSync(bool)
     [...snip...]
[30086:30239:0711/103542.012610:ERROR:ssl_client_socket_impl.cc(1108)] handshake failed; returned -1, SSL error code 1, net_error -107
[30086:30086:0711/103542.030871:WARNING:gaia_auth_fetcher.cc(949)] Could not reach Google Accounts servers: errno -107
BrowserTestBase received signal: Terminated. Backtrace:
#0 0x7f5e2ab9c08b base::debug::StackTrace::StackTrace()
#1 0x7f5e2ab9adcc base::debug::StackTrace::StackTrace()
#2 0x000005089a27 content::(anonymous namespace)::DumpStackTraceSignalHandler()
#3 0x7f5e0d3f8cb0 <unknown>
#4 0x7f5e0d4b2c5d __poll
#5 0x7f5e0ffa4fe4 <unknown>
#6 0x7f5e0ffa50ec g_main_context_iteration
#7 0x7f5e2ac4168f base::MessagePumpGlib::Run()
#8 0x7f5e2ac2f403 base::MessageLoop::Run()
#9 0x7f5e2accfe47 base::RunLoop::Run()
#10 0x000005127836 content::RunThisRunLoop()
#11 0x000005128491 content::MessageLoopRunner::Run()
#12 0x000005128aad content::WindowedNotificationObserver::Wait()
#13 0x000001a15f22 chromeos::BlockingLoginTest::Login()
#14 0x000001a14e5b chromeos::BlockingLoginTest_LoginBlocksForUser_Test::RunTestOnMainThread()
#15 0x000005089717 content::BrowserTestBase::ProxyRunTestOnMainThreadLoop()
     [...snip...]
#58 0x000003fca28a main
#59 0x7f5e0d3e3f45 __libc_start_main
#60 0x0000008d1be4 <unknown>
[315/335] BlockingLoginTestInstance/BlockingLoginTest.LoginBlocksForUser/7 (TIMED OUT)

I couldn't locate a suspect CL.

+atwilson@ - Since you've indicated that this test is fundamentally broken anyway (Issue 580537), could you take a look?  Either fix, triage, or disable?  Thanks.
 

Comment 1 by derat@chromium.org, Jul 13 2017

Cc: ejcaruso@chromium.org
(+Eric because I'm surprised to see MessagePumpGlib in this stack trace!)
I thought we figured out that we did use this as the MessageLoopForUI in Chrome on Chrome OS? This shouldn't be that surprising...

Comment 3 by derat@chromium.org, Jul 13 2017

Cc: steve...@chromium.org
No, Linux Chrome uses it as its MessageLoopForUI since it needs to display GTK dialogs. I don't know why it's being used here. I wouldn't expect use_glib to be set on Chrome-for-Chrome-OS builds, even just for testing.

Sorry, I'll stop talking about it here since it's a tangent and likely unrelated to the test failure.
Mergedinto: 580537
Status: Duplicate (was: Assigned)

Sign in to add a comment