> On Nov. 15, 2012, 4:21 p.m., Benjamin Hindman wrote: > > Have the segfault issues been resolved? > > Vinod Kone wrote: > not yet.
Wasn't able to repro the above again :/ Ran all master detector tests in a loop (~200 times) and the new expiration tests (~50 times) with NO segfaults! I'm not sure what was happening before to cause the SEGFAULTs and what changed now. Feel free to test this patch (in a loop) on your own before committing to the trunk. - Vinod ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/7842/#review13469 ----------------------------------------------------------- On Nov. 2, 2012, 10:29 p.m., Vinod Kone wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/7842/ > ----------------------------------------------------------- > > (Updated Nov. 2, 2012, 10:29 p.m.) > > > Review request for mesos, Benjamin Hindman and Ben Mahler. > > > Description > ------- > > Its quite hard to test true network partition behavior, so these tests expire > zk sessions as a close approximation. > > > Diffs > ----- > > src/detector/detector.cpp 57febbb22e2abf193f7881839b31db69d49b14df > src/tests/zookeeper_tests.cpp fff53b84356243663cfc75d66ddb3a1a7faf8043 > > Diff: https://reviews.apache.org/r/7842/diff/ > > > Testing > ------- > > make check. > > When running in a repeat loop, I found the following JVM error (in iteration > 12). It looks un-related to current tests > > [vinod@smfd-aki-27-sr1:~/mesos/build] $ GLOG_v=1 ./bin/mesos-tests.sh > --gtest_filter="*Expire*" --verbose --gtest_repeat=-1 > --gtest_break_on_failure > .... > .... > I1102 22:21:46.203616 28200 detector.cpp:331] Created ephemeral/sequence > znode at '/mesos/0000000003' > I1102 22:21:46.203902 28198 detector.cpp:486] Master detector > ((290)@10.35.12.124:57104) found 2 registered masters > I1102 22:21:46.204463 28200 detector.cpp:486] Master detector > ((287)@10.35.12.124:57104) found 2 registered masters > I1102 22:21:46.205245 28200 detector.cpp:532] Master detector > ((287)@10.35.12.124:57104) got new master pid: (290)@10.35.12.124:57104 > ../../src/tests/zookeeper_tests.cpp:441: Failure > Mock function called more times than expected - returning directly. > Function call: newMasterDetected(@0x451ffe90 (290)@10.35.12.124:57104) > Expected: to be called once > Actual: called twice - over-saturated and active > # > # A fatal error has been detected by the Java Runtime Environment: > # > # SIGSEGV (0xb) at pc=0x000000000075a304, pid=28178, tid=1159727424 > # > # JRE version: 7.0_10-b10 > # Java VM: Java HotSpot(TM) 64-Bit Server VM (23.6-b04 mixed mode linux-amd64 > compressed oops) > # Problematic frame: > # C [lt-mesos-tests+0x35a304] cgroups::removeCgroup(std::string const&, > std::string const&)+0x35a304 > # > # Failed to write core dump. Core dumps have been disabled. To enable core > dumping, try "ulimit -c unlimited" before starting Java again > # > # An error report file with more information is saved as: > # /home/vinod/mesos/build/hs_err_pid28178.log > # > # If you would like to submit a bug report, please visit: > # http://bugreport.sun.com/bugreport/crash.jsp > # > Aborted > > > Thanks, > > Vinod Kone > >
