I cannot reproduce the failures locally. I am disabling it.

+clamy: git blame thinks you wrote this test. Please fix and reenable, or reassign it.

Sorry! wrong bug! Disregard my last comment!

All the flakes are from the same CL. May be a false positive.

I'm going to watch to see if further flakes arise before doing anything.

There have been two more failures, so I'm disabling.

Stack trace:
Received signal 11 SEGV_MAPERR 000000000048
#0 0x000004c5bbac base::debug::StackTrace::StackTrace()
#1 0x000004c5b721 base::debug::(anonymous namespace)::StackDumpSignalHandler()
#2 0x7fa7bba76330 <unknown>
#3 0x0000018499a3 content::FrameTreeTest_FindFrames_Test::TestBody()
#4 0x000002d4dc96 testing::Test::Run()
#5 0x000002d4e740 testing::TestInfo::Run()
#6 0x000002d4ec27 testing::TestCase::Run()
#7 0x000002d5a517 testing::internal::UnitTestImpl::RunAllTests()
#8 0x000002d5a09e testing::UnitTest::Run()
#9 0x0000044b6b92 base::TestSuite::Run()
#10 0x0000044bf730 base::(anonymous namespace)::LaunchUnitTestsInternal()
#11 0x0000044bf580 base::LaunchUnitTests()
#12 0x000002441bd5 main
#13 0x7fa7b7f52f45 __libc_start_main
#14 0x0000014c602a _start
  r8: 0000000000000000  r9: 134be1f28e6dce31 r10: 37d64ae808bd0c89 r11: 00000f5026f9d2e0
 r12: 00000f5026ce3960 r13: 00000f5026c87200 r14: 00000f5026c87200 r15: 000000005abbd7f0
  di: 00000f5026eabea8  si: 0000000000000001  bp: 00007fff5ec11dd0  bx: 00000f5026fc7b90
  dx: 00007fa7bbe78c60  ax: 0000000000000000  cx: 0000000007c56988  sp: 00007fff5ec11c10
  ip: 00000000018499a3 efl: 0000000000010202 cgf: 0000000000000033 erf: 0000000000000004
 trp: 000000000000000e msk: 0000000000000000 cr2: 0000000000000048
[end of stack trace]

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1c854a22e9cf92fbbb86ac46ba6b56c4bac1764a

commit 1c854a22e9cf92fbbb86ac46ba6b56c4bac1764a
Author: Adam Rice <ricea@chromium.org>
Date: Thu Mar 29 13:42:31 2018

Disable test FrameTreeTest.FindFrames

Test FrameTreeTest.FindFrames is flaky. Disable it.

TBR=creis@chromium.org

Bug:  826599 
Change-Id: I3536ef001db9f8df4193ad7c709cc4f7c4246d97
Reviewed-on: https://chromium-review.googlesource.com/985996
Reviewed-by: Adam Rice <ricea@chromium.org>
Commit-Queue: Adam Rice <ricea@chromium.org>
Cr-Commit-Position: refs/heads/master@{#546808}
[modify] https://crrev.com/1c854a22e9cf92fbbb86ac46ba6b56c4bac1764a/content/browser/frame_host/frame_tree_unittest.cc

+creis git blame thinks you wrote this test. Please deflake & re-enable, or reassign this issue.

I don't think these failures are due to this (very old) test.  There were similar failures in other tests (e.g., FrameTreeTest.ProcessCrashClearsGlobalMap, FrameTreeNodeBlameContextTest.URLChange, MediaStreamUIProxyTest.DeleteBeforeAccepted), and that suggests some other cause is to blame.  Can you revert r546808 and take another look at when these tests started failing to see if something bigger is a problem?

For example, see the output in:
https://ci.chromium.org/p/chromium/builders/luci.chromium.try/linux_chromium_rel_ng/56305

I'll revert in https://chromium-review.googlesource.com/c/chromium/src/+/990752 if it passes try jobs.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/b140562cdcac6b2f7d34c95b0a1d43e6add0fd83

commit b140562cdcac6b2f7d34c95b0a1d43e6add0fd83
Author: Charlie Reis <creis@chromium.org>
Date: Mon Apr 02 22:52:39 2018

Revert "Disable test FrameTreeTest.FindFrames"

This reverts commit 1c854a22e9cf92fbbb86ac46ba6b56c4bac1764a.

Reason for revert: Failures were not due to this test.  Passes locally.

Original change's description:
> Disable test FrameTreeTest.FindFrames
> 
> Test FrameTreeTest.FindFrames is flaky. Disable it.
> 
> TBR=creis@chromium.org
> 
> Bug:  826599 
> Change-Id: I3536ef001db9f8df4193ad7c709cc4f7c4246d97
> Reviewed-on: https://chromium-review.googlesource.com/985996
> Reviewed-by: Adam Rice <ricea@chromium.org>
> Commit-Queue: Adam Rice <ricea@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#546808}

TBR=creis@chromium.org,ricea@chromium.org

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug:  826599 
Change-Id: I1b3aeeaccb8458fc57959efdf236d126280cde5b
Reviewed-on: https://chromium-review.googlesource.com/990752
Reviewed-by: Charlie Reis <creis@chromium.org>
Commit-Queue: Charlie Reis <creis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#547545}
[modify] https://crrev.com/b140562cdcac6b2f7d34c95b0a1d43e6add0fd83/content/browser/frame_host/frame_tree_unittest.cc

#8 The issue is not whether or not the test is to blame for the flakiness. The issue is that the flaky failures of this test hurt the velocity of all Chrome developers by making the CQ slower and less reliable.

Until the root cause of the flakiness can be found and fixed, the test should remain disabled.

#9: I totally agree when there's a single test failing in a flaky way, or a small number of related tests that seem like they're facing the same problem.  Disabling it to get things green is the right way to unblock everyone.

However, when a huge list of unrelated tests have all started failing at once, I strongly disagree that we should just disable them all.  That's more indicative of either an independent regression that needs to be found right away, or a bot issue that troopers should help with.  Disabling a huge list of tests in such a situation would harm our overall test coverage and increases the risk of regressions.

I understand that it may not have been immediately evident that this was a case where the test wasn't to blame, but once this was established, re-enabling the test was the right thing to do.  (There don't appear to be new occurrences of the failure.)