> On July 23, 2015, 11:51 p.m., Vinod Kone wrote: > > src/tests/master_tests.cpp, line 1022 > > <https://reviews.apache.org/r/36663/diff/2/?file=1020111#file1020111line1022> > > > > This test is testing the createMasterInfo() method which is not used in > > production. Can you also write a test with ZK in play? That is going to > > test the changes you made in master.cpp.
Yep - I had looked in the zookeeper_* tests, but found nothing related to MasterInfo; turns out the tests are in the master_contender.cpp tests. Added one there. - Marco ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/36663/#review92834 ----------------------------------------------------------- On July 23, 2015, 11:54 p.m., Marco Massenzio wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/36663/ > ----------------------------------------------------------- > > (Updated July 23, 2015, 11:54 p.m.) > > > Review request for mesos, Anand Mazumdar and Vinod Kone. > > > Bugs: MESOS-2736 > https://issues.apache.org/jira/browse/MESOS-2736 > > > Repository: mesos > > > Description > ------- > > Added address field to MasterInfo > > As part of the new HTTP API and the need to > provide a better interface for clients that > do not integrate libmesos, we provide the IP > address of the Leader Master in the information > that gets serialized (in JSON) to ZooKeeper. > > This will eventually supersede the `ip`, `port` > and `hostname` fields that are currently in > MasterInfo an which cannot fully support IPv6 > addresses. > > Jira: MESOS-2736 > > Review: https://reviews.apache.org/r/36663 > > > Diffs > ----- > > include/mesos/mesos.proto e015c81d5052214ef8207642e23b3892a6123c9a > include/mesos/scheduler/scheduler.proto > 89daf8a6b74057ee156b3ad691397e76fcb835b8 > src/common/protobuf_utils.cpp e0f82b53f5e106bbf4e21d6ac946df0fae821882 > src/master/master.hpp bf61bb2deb61a73303f4122383dcb7e8f5d726de > src/master/master.cpp bab04feaf238e51c4d50feb1d3a9c0735ec68fec > src/scheduler/scheduler.cpp badc107dbf4e2bd9146f9244724e0db5c2ae05d3 > src/tests/master_tests.cpp 826f276317da2507cdf99b2f2e21a147444469d8 > src/tests/scheduler_tests.cpp 98fc70bf43ba99b54064a236795c7e1269004b71 > > Diff: https://reviews.apache.org/r/36663/diff/ > > > Testing > ------- > > make check > (also tested via [zk-mesos](https://github.com/massenz/zk-mesos) that the > information serialized to ZK is readable and as expected). > > Also ran 2x 0.23 master builds against a 0.24 one with this patch applied; > getting both versions in turn to be Leader, and also ran a 0.23 Slave with a > 0.24 Leader, and they all recognized each other. > > > Thanks, > > Marco Massenzio > >
