----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/13087/#review28019 -----------------------------------------------------------
src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54478> reorder. src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54479> new line? src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54481> The test teardown should delete the work directory. So you don't need to do this explicitly. src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54483> Why not use the/a cluster abstraction? src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54484> put url.get().servers on the next line. Also, maybe we should've a Group constructor that takes zookeeper::URL. src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54485> pull this down to where it's used. src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54486> pull this down to where it's used. src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54487> Why are these 2 timeouts different? src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54488> ditto. pull down. src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54489> s/down/down ZooKeeper/ src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54493> s/don/do/ src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54499> Why not just use the default timeout? src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54495> pull down. src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54496> s/contender1/followerContender/? src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54497> kill new line. src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54498> Inducing lost connections this way is a bit weird because the connections are actually alive on the ZKServer. Can you instead just bring down ZK to achieve the same thing? src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54501> Reorder. src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54502> Also check the new leader is the 'follower'? src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54503> Instead of doing a sleep can you AWAIT on slave reconnecting with ZK? src/tests/master_contender_detector_tests.cpp <https://reviews.apache.org/r/13087/#comment54504> Isn't 10s the default? - Vinod Kone On Oct. 30, 2013, 5:15 p.m., Jiang Yan Xu wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/13087/ > ----------------------------------------------------------- > > (Updated Oct. 30, 2013, 5:15 p.m.) > > > Review request for mesos, Benjamin Hindman, Ben Mahler, and Vinod Kone. > > > Bugs: MESOS-496 > https://issues.apache.org/jira/browse/MESOS-496 > > > Repository: mesos-git > > > Description > ------- > > See summary. > > > Diffs > ----- > > include/mesos/scheduler.hpp fa1ffe8b24a054556593d31eeae3fd1ac1761fb5 > src/Makefile.am a11c76b998ccc84daca02facddd20a4acb88aa43 > src/cli/resolve.cpp b05f5103ca898f74acb7dfe8f423ffc7e1872e09 > src/detector/detector.hpp 3aaebfe89f0b4bb5c64e7f4c8974d8b13bc6f6b1 > src/detector/detector.cpp 8d9f118757192ca38bcdb26663ed86effbcf3c9e > src/local/local.cpp 180756a929baed4dce97fdb8774f07010587468f > src/local/main.cpp 5995c538077dea301f231668b4b5f905e73e7b3b > src/master/contender.hpp PRE-CREATION > src/master/contender.cpp PRE-CREATION > src/master/detector.hpp PRE-CREATION > src/master/detector.cpp PRE-CREATION > src/master/http.cpp f2a535a002f07784db724f6fdedfb26b271d6327 > src/master/main.cpp 45caf9dde0af5995dde73ae82e83e1d7169d17b8 > src/master/master.hpp e377af8b3ccd932ae411fa2df4c19642a7310d02 > src/master/master.cpp 8e14a070e87ebe579b54d05fb1e8b286edb5e459 > src/messages/messages.proto a5dded2c0f3f88a5f7d789ccde6e59f68880374b > src/sched/sched.cpp 3049096a7aeaea6c1063a1bbcef119517daa236c > src/slave/http.cpp 62fbb37a1924062543bf9db4229704bdef91601d > src/slave/main.cpp 750a12766bde64059bfd4635ea077cbd43cb4301 > src/slave/slave.hpp 68526f3fb01be6dc19d7e3c8e4ba8220b2c56ccb > src/slave/slave.cpp 74bcbecb8240a4de21efd103be744869a81c72ce > src/tests/allocator_tests.cpp b0beb72273e9d44a407d0e83e6c18a541b212089 > src/tests/authentication_tests.cpp 48a9323d03416ad9ee25fc19838d89678ff613bc > src/tests/cluster.hpp bea395a73b09df287fef743dae1aa836f4dac691 > src/tests/fault_tolerance_tests.cpp > 3989e14eb926d7f9b5c7731aa0713a56fc3fb6e9 > src/tests/gc_tests.cpp 5459b78126d3607114ea171d3f76883f8355751c > src/tests/isolator_tests.cpp ab5b00aa7f12dbd299debc1a405ac50c59cec85c > src/tests/master_contender_detector_tests.cpp PRE-CREATION > src/tests/master_detector_tests.cpp > 06c586d29996cd599ae58dfd49cf1324aac0a6d6 > src/tests/master_tests.cpp bf790d24ff6b2fb199a04caebef523797b075e63 > src/tests/mesos.hpp ad39c095f9699fbe7958d52698fd2b0319e84eb9 > src/tests/mesos.cpp 351ffd614250847c52657b721a560eb2cf7c9443 > src/tests/slave_recovery_tests.cpp 37faf2054750080902cb06b62da9a9e2252c566f > > Diff: https://reviews.apache.org/r/13087/diff/ > > > Testing > ------- > > make check 100 times > > > Thanks, > > Jiang Yan Xu > >
