[MESOS-3773] RegistryClientTest.SimpleGetBlob is flaky

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Fix Version/s: 0.24.2, 0.25.1, 0.26.1, 0.27.0
• Component/s: test
• Labels: mesosphere
• Sprint: Mesosphere Sprint 21, Mesosphere Sprint 24
• Story Points: 4
• Assignee: Jojy Varghese
• Reporter: Joseph Wu
• Shepherd: Joris Van Remoortere
• Created: 20/Oct/15 21:03
• Updated: 26/Nov/18 12:52
• Resolved: 26/Nov/18 12:52

Description

RegistryClientTest.SimpleGetBlob fails about 1/5 times. This was encountered on OSX.

Repro

bin/mesos-tests.sh --gtest_filter="*RegistryClientTest.SimpleGetBlob*" --gtest_repeat=10 --gtest_break_on_failure

Example Failure

[ RUN      ] RegistryClientTest.SimpleGetBlob
../../src/tests/containerizer/provisioner_docker_tests.cpp:946: Failure
Value of: blobResponse
  Actual: "2015-10-20 20:58:59.579393024+00:00"
Expected: blob.get()
Which is: "\x15\x3\x3\00(P~\xCA&\xC6<\x4\x16\xE\xB2\xFF\b1a\xB9Z{\xE0\x80\xDA`\xBCt\x5R\x81x6\xF8 \x8B{\xA8\xA9\x4\xAB\xB6" "E\xE6\xDE\xCF\xD9*\xCC!\xC2\x15" "2015-10-20 20:58:59.579393024+00:00"
*** Aborted at 1445374739 (unix time) try "date -d @1445374739" if you are using GNU date ***
PC: @        0x103144ddc testing::UnitTest::AddTestPartResult()
*** SIGSEGV (@0x0) received by PID 49008 (TID 0x7fff73ca3300) stack trace: ***
    @     0x7fff8c58af1a _sigtramp
    @     0x7fff8386e187 malloc
    @        0x1031445b7 testing::internal::AssertHelper::operator=()
    @        0x1030d32e0 mesos::internal::tests::RegistryClientTest_SimpleGetBlob_Test::TestBody()
    @        0x1030d3562 mesos::internal::tests::RegistryClientTest_SimpleGetBlob_Test::TestBody()
    @        0x1031ac8f3 testing::internal::HandleSehExceptionsInMethodIfSupported<>()
    @        0x103192f87 testing::internal::HandleExceptionsInMethodIfSupported<>()
    @        0x1031533f5 testing::Test::Run()
    @        0x10315493b testing::TestInfo::Run()
    @        0x1031555f7 testing::TestCase::Run()
    @        0x103163df3 testing::internal::UnitTestImpl::RunAllTests()
    @        0x1031af8c3 testing::internal::HandleSehExceptionsInMethodIfSupported<>()
    @        0x103195397 testing::internal::HandleExceptionsInMethodIfSupported<>()
    @        0x1031639f2 testing::UnitTest::Run()
    @        0x1025abd41 RUN_ALL_TESTS()
    @        0x1025a8089 main
    @     0x7fff86b155c9 start

Less common failure

[ RUN      ] RegistryClientTest.SimpleGetBlob
../../src/tests/containerizer/provisioner_docker_tests.cpp:926: Failure
(socket).failure(): Failed accept: connection error: error:00000000:lib(0):func(0):reason(0)

Issue Links

• duplicates: MESOS-3807 RegistryClientTest.SimpleGetManifest is flaky
• duplicates: MESOS-4202 Race in SSL socket shutdown
• is duplicated by: MESOS-3798 io::write(fd, const string&) api writes junk sometimes
• is duplicated by: MESOS-4026 RegistryClientTest.SimpleRegistryPuller is flaky
• is related to: MESOS-4026 RegistryClientTest.SimpleRegistryPuller is flaky
• relates to: MESOS-3726 RegistryClientTest.SimpleGetBlob is flaky
• relates to: MESOS-3798 io::write(fd, const string&) api writes junk sometimes
• relates to: MESOS-4057 RegistryClientTest suite fails reliably in optimized build