----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/70788/#review215712 -----------------------------------------------------------
Fix it, then Ship it! src/tests/api_tests.cpp Lines 302 (patched) <https://reviews.apache.org/r/70788/#comment302519> What do you think of explicitly asserting what RP state each of these `UpdateSlaveMessage` contains, e.g., here (using default message if unset) ``` ASSERT_EQ(0, updateSlaveMessage->resource_providers().providers_size()); ``` Similarly for each instance below. - Benjamin Bannier On June 6, 2019, 6:56 a.m., Chun-Hung Hsiao wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/70788/ > ----------------------------------------------------------- > > (Updated June 6, 2019, 6:56 a.m.) > > > Review request for mesos and Benjamin Bannier. > > > Bugs: MESOS-9831 > https://issues.apache.org/jira/browse/MESOS-9831 > > > Repository: mesos > > > Description > ------- > > Previously when there is a missing resource provider in an > `UpdateSlaveMessage` but the agent's total resources remain unchanged, > the update message will be completely ignored, so the missing resource > provider will still be cached in the master's state, which is not the > desired behavior. This patch ensures that the master's state gets > updated if any resource provider is missing. > > > Diffs > ----- > > src/master/master.cpp 4d7c37cf8814e45432b3fe15173f5343676a372b > src/tests/api_tests.cpp 37d0cb11ddb24b32f1e83feefe168f2aa2eb659c > > > Diff: https://reviews.apache.org/r/70788/diff/1/ > > > Testing > ------- > > make check > > > Thanks, > > Chun-Hung Hsiao > >
