media_router integration tests failing on Linux buildbots |
|||||||
Issue descriptionmedia_router tests are failing on the Linux bots. It looks like a problem with the test script. I didn't see the problematic code at a quick look. Running it in a browser and inspecting the page should should narrow down the culprit. [ RUN ] MediaRouterIntegrationBrowserTest.MANUAL_Dialog_RouteCreationTimedOut Xlib: extension "RANDR" missing on display ":9". Xlib: extension "RANDR" missing on display ":9". [5373:5373:0214/061030.867632:WARNING:audio_manager.cc(322)] Multiple instances of AudioManager detected [5373:5373:0214/061030.867659:WARNING:audio_manager.cc(279)] Multiple instances of AudioManager detected [5373:5373:0214/061030.892299:WARNING:password_store_factory.cc(247)] Using basic (unencrypted) store for password storage. See https://chromium.googlesource.com/chromium/src/+/master/docs/linux_password_storage.md for more information about password storage options. [5373:5373:0214/061031.444769:ERROR:CONSOLE(1)] "Uncaught ReferenceError: media_router is not defined", source: (1) BrowserTestBase received signal: Terminated. Backtrace: #0 0x0000027e0007 base::debug::StackTrace::StackTrace() #1 0x000002d38366 content::(anonymous namespace)::DumpStackTraceSignalHandler() #2 0x7fd125eeecb0 <unknown> #3 0x7fd125fa4fdd __poll #4 0x7fd12b950fe4 <unknown> #5 0x7fd12b9510ec g_main_context_iteration #6 0x0000027f9c36 base::MessagePumpGlib::Run() #7 0x0000027f7dc7 base::MessageLoop::RunHandler() #8 0x0000028173ae base::RunLoop::Run() #9 0x0000014b5594 media_router::MediaRouterBaseBrowserTest::ConditionalWait() #10 0x0000014ba8a9 media_router::MediaRouterIntegrationBrowserTest::WaitUntilIssue() #11 0x0000014c3865 media_router::MediaRouterIntegrationBrowserTest_MANUAL_Dialog_RouteCreationTimedOut_Test::RunTestOnMainThread() #12 0x00000288234f InProcessBrowserTest::RunTestOnMainThreadLoop() #13 0x000002d380c8 content::BrowserTestBase::ProxyRunTestOnMainThreadLoop() #14 0x000002bb5827 ChromeBrowserMainParts::PreMainMessageLoopRunImpl() #15 0x000002bb48f1 ChromeBrowserMainParts::PreMainMessageLoopRun() #16 0x0000018ece35 content::BrowserMainLoop::PreMainMessageLoopRun() #17 0x000001baf886 content::StartupTaskRunner::RunAllTasksNow() #18 0x0000018eafeb content::BrowserMainLoop::CreateStartupTasks() #19 0x0000018f061b content::BrowserMainRunnerImpl::Initialize() #20 0x0000018e8638 content::BrowserMain() #21 0x0000027cb02c content::ContentMainRunnerImpl::Run() #22 0x0000027c9bb0 content::ContentMain() #23 0x000002d379f9 content::BrowserTestBase::SetUp() #24 0x000002880a80 InProcessBrowserTest::SetUp() #25 0x00000331334e testing::Test::Run() #26 0x000003313f10 testing::TestInfo::Run() #27 0x000003314287 testing::TestCase::Run() #28 0x00000331b367 testing::internal::UnitTestImpl::RunAllTests() #29 0x00000331afe7 testing::UnitTest::Run() #30 0x000002891a21 base::TestSuite::Run() #31 0x0000027d4788 ChromeTestSuiteRunner::RunTestSuite() #32 0x000002d5f4ea content::LaunchTests() #33 0x0000027d4707 main #34 0x7fd125ed9f45 __libc_start_main #35 0x0000005f7391 <unknown> [12/12] MediaRouterIntegrationBrowserTest.MANUAL_Dialog_RouteCreationTimedOut (TIMED OUT) clifford is this addressed by your recent changes?
,
Feb 14 2017
I can repro locally on Linux. I will add some logging in the test to see why it failed.
,
Feb 15 2017
Hmm, maybe the failure was a one-off, as I can't repro it anymore. I suspect the Linux machine might've gotten really slow. You can see that the other tests are also timing out occasionally but pass on a retry. The RouteCreationTimedOut test always times out because it is one of the longest tests. We need to be able to run the tests on a faster machine. Barring that, we will need to increase the browser test timeouts.
,
Feb 15 2017
The only thing that I worry is the error message "Uncaught ReferenceError: media_router is not defined" I also tried to run it locally with the latest Chrome and last extension, but can't repro the issue either. There is 1 time that needs to retry, but I don't see that message. The integration are all running on VM and it's not very feasible to run on a faster machine. Can you increase the wait time for MediaRouterIntegrationBrowserTest::WaitUntilIssue() and see if it gets better?
,
Feb 16 2017
,
Feb 16 2017
Chatted with Clifford about this. The Uncaught ReferenceError is seemingly unrelated but we should look into this separately to reduce noise in our tests. MANUAL_Dialog_RouteCreationTimedOut timed out because it has exceeded the browser test time limit which is 30s by default. So increasing the time limit by passing in a --ui-test-action-max-timeout switch in the Media Router recipe would be a reasonable stop-gap (I noticed other tests were also timed out flaky, so increasing the timeout across the board should also help with that). If that doesn't help, we should reach out to infra team to help us investigate the source of slow down. It is not reasonable to keep increasing test timeout indefinitely. Separately, I will look into whether we can inject a smaller route creation timeout for this test case, since it typically takes 28-29s for this test to complete locally on my machine, which is cutting it real close.
,
Feb 17 2017
After changing the timeout to 60s, the integration tests on linux are passing in the latest run. MANUAL_Dialog_RouteCreationTimedOut takes 43522ms to finish in the latest run, which is more than the default 30s timeout. Mark this as fixed now.
,
Feb 17 2017
|
|||||||
►
Sign in to add a comment |
|||||||
Comment 1 by cliffordcheng@chromium.org
, Feb 14 2017Owner: imch...@chromium.org