----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/52408/#review150956 -----------------------------------------------------------
src/slave/slave.hpp (line 328) <https://reviews.apache.org/r/52408/#comment219064> s/activateExecutorAPiEndpoint/allowExecutorSubscription/ . If we are looking for shorter name we can call it "reconcile" or "reconnect" which is what the comment above suggests. that way it applies to both PID and HTTP based executors. we can also shutdown pid executors if they try to re-register when this flag is false; shouldn't happen if they are using the driver but there are executors which speak pid protocol without using the driver. - Vinod Kone On Sept. 29, 2016, 10:08 p.m., Anand Mazumdar wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/52408/ > ----------------------------------------------------------- > > (Updated Sept. 29, 2016, 10:08 p.m.) > > > Review request for mesos and Vinod Kone. > > > Bugs: MESOS-6274 > https://issues.apache.org/jira/browse/MESOS-6274 > > > Repository: mesos > > > Description > ------- > > Previously, it was possible for a HTTP based executor to subscribe > with the agent before the containerizer recovery is done. This > was a problem since calling `containerizer->update()` etc. would > result in a failure. > > > Diffs > ----- > > src/slave/http.cpp ff750e23433ce241d76c3b163b8442f9bd4657b1 > src/slave/slave.hpp 7b414d06ffac6db8434ba5c4e7be1989f2ca0864 > src/slave/slave.cpp 46bc8b38cf38c20b01532455b4d369221775c837 > src/tests/executor_http_api_tests.cpp > 14a907f1955513fbcd46af2fc65c1ce5c85bf8cb > > Diff: https://reviews.apache.org/r/52408/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Anand Mazumdar > >
