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

Issue 813631 link

Starred by 3 users

Issue metadata

Status: Assigned
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Fuchsia
Pri: 2
Type: Bug

Blocking:
issue 738275



Sign in to add a comment

URLRequestQuicTest.CancelPushIfCached_AllCached/SomeCached flaked in net_unittests on Fuchsia

Project Member Reported by scottmg@chromium.org, Feb 19 2018

Issue description

https://build.chromium.org/p/chromium.fyi/builders/Fuchsia/builds/14138

https://chromium-swarm.appspot.com/user/task/3bc72700872d8210

[00293.930] 03945.03981> [ RUN      ] URLRequestQuicTest.CancelPushIfCached_AllCached
[00293.930] 03945.03981> ../../net/url_request/url_request_quic_unittest.cc:436: Failure
[00293.930] 03945.03981> Expected: (2u) <= (GetRstErrorCountReceivedByServer(QUIC_STREAM_CANCELLED)), actual: 2 vs 0
[00293.931] 03945.03981> [  FAILED  ] URLRequestQuicTest.CancelPushIfCached_AllCached (2852 ms)

This fixture was noted flaky in https://bugs.chromium.org/p/chromium/issues/detail?id=763098 but don't think this looks the same.
 
Project Member

Comment 1 by bugdroid1@chromium.org, Feb 19 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/33b9ef7d1a281a80f4f3d81f3c09bf6b116c5c43

commit 33b9ef7d1a281a80f4f3d81f3c09bf6b116c5c43
Author: Scott Graham <scottmg@chromium.org>
Date: Mon Feb 19 19:09:39 2018

fuchsia: Disable URLRequestQuicTest.CancelPushIfCached_AllCached

Flaked at
https://build.chromium.org/p/chromium.fyi/builders/Fuchsia/builds/14138.

TBR: wez@chromium.org
Bug: 813631
Change-Id: Ifda5f9f3efc04fa8f17859678713b9658e3652d7
Reviewed-on: https://chromium-review.googlesource.com/925450
Commit-Queue: Scott Graham <scottmg@chromium.org>
Reviewed-by: Scott Graham <scottmg@chromium.org>
Cr-Commit-Position: refs/heads/master@{#537678}
[modify] https://crrev.com/33b9ef7d1a281a80f4f3d81f3c09bf6b116c5c43/testing/buildbot/filters/fuchsia.net_unittests.filter

Comment 2 by w...@chromium.org, Feb 20 2018

Cc: rch@chromium.org
+rch@, do you know what the nature of the failure is that this check is catching?  Is it most likely that we're getting RSTs of some different sort?

Should we restructure this test code to use more explicit expectations and/or to print what RSTs _were_ handled, in cases like this?
It looks like the test wanted to confirm that server sent two RSTs to the client and client received the two RSTs with error code QUIC_STREAM_CANCELLED. 

Looks like after "some change", the tests exits before client received the two RSTs. 
Status: Untriaged (was: Unconfirmed)

Comment 5 by w...@chromium.org, Apr 27 2018

Cc: jam...@chromium.org w...@chromium.org
 Issue 837539  has been merged into this issue.
Project Member

Comment 6 by bugdroid1@chromium.org, Apr 27 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/366f1bc86ba9a876dbc1c5e537bc42adb510fea9

commit 366f1bc86ba9a876dbc1c5e537bc42adb510fea9
Author: Wez <wez@chromium.org>
Date: Fri Apr 27 14:03:42 2018

Broaden Fuchsia filter on URLRequestQuicTest.CancelPushIfCached_* tests.

Bug: 813631
Change-Id: I4d9bfa6b271129b085f1b3c6bf376bf68a24cd59
Reviewed-on: https://chromium-review.googlesource.com/1032262
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Wez <wez@chromium.org>
Cr-Commit-Position: refs/heads/master@{#554377}
[modify] https://crrev.com/366f1bc86ba9a876dbc1c5e537bc42adb510fea9/testing/buildbot/filters/fuchsia.net_unittests.filter

Summary: URLRequestQuicTest.CancelPushIfCached_AllCached/SomeCached flaked in net_unittests on Fuchsia (was: URLRequestQuicTest.CancelPushIfCached_AllCached flaked in net_unittests on Fuchsia)
Also seeing flakes for URLRequestQuicTest.CancelPushIfCached_SomeCached.
https://logs.chromium.org/v/?s=chromium%2Fbb%2Fchromium.fyi%2FFuchsia%2F16607%2F%2B%2Frecipes%2Fsteps%2Fnet_unittests%2F0%2Flogs%2FURLRequestQuicTest.CancelPushIfCached_SomeCached%2F0

Comment 8 by w...@chromium.org, Apr 27 2018

Isn't that the one I filed  issue 837539  for? That's why I broadened the filter.
Cc: scottmg@chromium.org
URLRequestQuicTest.DoNotCancelPushIfNotFoundInCache failed at
https://logs.chromium.org/v/?s=chromium%2Fbuildbucket%2Fcr-buildbucket.appspot.com%2F8935416162173798272%2F%2B%2Fsteps%2Fnet_unittests%2F0%2Fstdout

I will disable it.


[00144.369] devmgr: crash_analyzer_listener: analyzing exception type 0x108
[00144.369] <== fatal exception: process /pkg/net_unittests__exec[302556] thread initial-thread[302573]
[00144.369] <== fatal page fault, PC at 0x7c468481fd2
[00144.369]  CS:                   0 RIP:      0x7c468481fd2 EFL:            0x10246 CR2:               0x20
[00144.369]  RAX:                  0 RBX:     0x6d845b724598 RCX:                  0 RDX:                  0
[00144.370]  RSI:                  0 RDI:     0x6d845b724600 RBP:                0x3 RSP:     0x6d845b724550
[00144.370]   R8:               0x48  R9:         0x12e5412c R10:                  0 R11:              0x246
[00144.370]  R12:     0x6d845b724560 R13:      0x12e54193100 R14:      0x12e54193100 R15:                  0
[35321
00144.370]  errc:               0x4
[00144.370] bottom of user stack:
[00144.370] 0x00006d845b724550: 5b724650 00006d84 5b724768 00006d84 |PFr[.m..hGr[.m..|
[00144.370] 0x00006d845b724560: 9eec2d00 000001ad 0000003b 00000000 |.-......;.......|
[00144.370] 0x00006d845b724570: 00000080 80000000 00000020 00000000 |........ .......|
[00144.370] 0x00006d845b724580: 9eec4a10 000001ad 00000000 00000000 |.J..............|
[00144.370] 0x00006d845b724590: 00000000 00000000 00000000 00000000 |................|
[00144.370] 0x00006d845b7245a0: 00000000 00000000 9eec5820 000001ad |........ X......|
[00144.370] 0x00006d845b7245b0: 00000024 00000000 00000030 80000000 |$.......0.......|
[00144.370] 0x00006d845b7245c0: 9eec4b90 000001ad 00000025 00000000 |.K......%.......|
[00144.370] 0x00006d845b7245d0: 00000030 80000000 00000000 00000000 |0...............|
[00144.370] 0x00006d845b7245e0: 00000000 00000000 00000000 00000000 |................|
[00144.370] 0x00006d845b7245f0: 00000000 00000000 dd840004 80000000 |................|
[00144.370] 0x00006d845b724600: 00000000 00000000 00000000 00000000 |................|
[00144.370] 0x00006d845b724610: 00000000 00000000 00000000 00000000 |................|
[00144.370] 0x00006d845b724620: 00000000 00000000 00000000 ffffffff |................|
[00144.370] 0x00006d845b724630: 00000000 ffffffff 00000008 00000010 |................|
[00144.370] 0x00006d845b724640: 00000000 ffffffff 00000018 0000000c |................|
[00144.370] arch: x86_64
[00144.435] dso: id=d7044f42a317d302fab405001c538b4281ff92e1 base=0x585f3822d000 name=<vDSO>
[00144.435] dso: id=eb44c3bfb2ca4616f1b615d3f224548465be95a9 base=0x4d3154445000 name=libc.so
[00144.435] dso: id=dfce0b7e2cfbe5bf base=0x7c4668fa000 name=app:/pkg/net_unittests__exec
[00144.435] dso: id=5699979afa052b6b3d6522df994048490961519e base=0x34de4f57000 name=libfdio.so
[00144.435] dso: id=d3d232011d2079c0db3d064a9da37b7b4985bc9a base=0x2dd56442000 name=libasync-default.so
#01: net::URLRequestQuicTest_DoNotCancelPushIfNotFoundInCache_Test::TestBody() at url_request_quic_unittest.cc:?
#02: testing::Test::Run() at gtest.cc:?
#03: testing::TestInfo::Run() at gtest.cc:?
#04: testing::TestCase::Run() at gtest.cc:?
#05: testing::internal::UnitTestImpl::RunAllTests() at gtest.cc:?
#06: testing::UnitTest::Run() at gtest.cc:?
#07: base::TestSuite::Run() at test_suite.cc:?
#08: int base::internal::Invoker<base::internal::BindState<int (base::TestSuite::*)(), base::internal::UnretainedWrapper<NetTestSuite> >, int ()>::RunImpl<int (base::TestSuite::* const&)(), std::__2::tuple<base::internal::UnretainedWrapper<NetTestSuite> > const&, 0ul>(int (base::TestSuite::* const&)(), std::__2::tuple<base::internal::UnretainedWrapper<NetTestSuite> > const&, std::__2::integer_sequence<unsigned long, 0ul>) at run_all_unittests.cc:?
#09: base::OnceCallback<bool ()>::Run() && at host_resolver_impl.cc:?
#10: base::(anonymous namespace)::LaunchUnitTestsInternal(base::OnceCallback<int ()>, unsigned long, int, bool, base::OnceCallback<void ()>) at unit_test_launcher.cc:?
#11: base::LaunchUnitTests(int, char**, base::OnceCallback<int ()>) at unit_test_launcher.cc:?
#12: main at run_all_unittests.cc:?
#13: pc 0x4d315445eb8f sp 0x6d845b724fe0 (libc.so,0x19b8f)
#14: pc 0 sp 0x6d845b725000
[00144.445] {{{bt: 0x7c468481fd2}}}
[00144.449] {{{bt: 0x7c468723ddb}}}
[00144.450] {{{bt: 0x7c4687244a4}}}
[00144.450] {{{bt: 0x7c4687247dc}}}
[00144.450] {{{bt: 0x7c46872aa07}}}
[00144.451] {{{bt: 0x7c46872a7bd}}}
[00144.765] {{{bt: 0x7c46933b52e}}}
[00144.794] {{{bt: 0x7c4682811e8}}}
[00144.794] {{{bt: 0x7c4687089ac}}}
[00144.795] {{{bt: 0x7c46933e2b3}}}
[00144.795] {{{bt: 0x7c46933e165}}}
[00144.796] {{{bt: 0x7c4682810b6}}}
[00144.796] {{{bt: 0x4d315445eb8f}}}
[00144.796] {{{bt: 0}}}
Was the flakiness introduced recently? The tests were added more than a year ago, if the flakiness was introduced recently, it feels more like a regression in code or infra. 
Yes, this particular test seems to have started recently, but related tests (URLRequestQuicTest.CancelPushIfCached_*) were flaking back in February, so I was just wondering if the stack might say something obvious to you.
Project Member

Comment 13 by bugdroid1@chromium.org, Sep 14

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/f950deaf1ddd568940b9a520474c879e7f0f2f7f

commit f950deaf1ddd568940b9a520474c879e7f0f2f7f
Author: Scott Graham <scottmg@chromium.org>
Date: Fri Sep 14 18:18:11 2018

fuchsia: Disable URLRequestQuicTest.DoNotCancelPushIfNotFoundInCache due to flake

Bug: 813631
Change-Id: Ib179df82faf2052e86a7c1a26994eff4d9f0ab45
Reviewed-on: https://chromium-review.googlesource.com/1226355
Reviewed-by: Wez <wez@chromium.org>
Commit-Queue: Scott Graham <scottmg@chromium.org>
Cr-Commit-Position: refs/heads/master@{#591401}
[modify] https://crrev.com/f950deaf1ddd568940b9a520474c879e7f0f2f7f/testing/buildbot/filters/fuchsia.net_unittests.filter

Owner: w...@chromium.org
Status: Started (was: Untriaged)

Comment 15 by w...@chromium.org, Jan 20 (3 days ago)

Labels: -Pri-3 M-73 Pri-2
Owner: rch@chromium.org
Status: Assigned (was: Started)
Ryan, could you provide some pointers on how we should go about diagnosing why these three tests fail?

Sign in to add a comment