> On July 23, 2015, 6:24 p.m., Ben Mahler wrote: > > include/mesos/scheduler/scheduler.proto, lines 46-52 > > <https://reviews.apache.org/r/36708/diff/2/?file=1019318#file1019318line46> > > > > "does not receive heartbeats, or any other events, for an extended"? > > > > Just thinking that we don't care about heartbeats when there are lots > > of events flowing (and this may skew the scheduler's perception of time > > between heartbeats).
Wondering if we should also mention about the usage of a backoff strategy here ? Something like: s/resubscribe./resubscribe using a suitable backoff strategy. ? - Anand ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/36708/#review92784 ----------------------------------------------------------- On July 23, 2015, 12:23 a.m., Vinod Kone wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/36708/ > ----------------------------------------------------------- > > (Updated July 23, 2015, 12:23 a.m.) > > > Review request for mesos, Anand Mazumdar, Benjamin Hindman, and Ben Mahler. > > > Bugs: MESOS-3131 > https://issues.apache.org/jira/browse/MESOS-3131 > > > Repository: mesos > > > Description > ------- > > This is to deal with network partition and also making sure that network > intermediaries do not close the persistent connection between scheduler and > master. > > Will add the code in the master once it can speak HTTP. > > > Diffs > ----- > > include/mesos/scheduler/scheduler.proto > e2ca8e5cfb3d6405bf7be93d621743ce48c6c7e6 > > Diff: https://reviews.apache.org/r/36708/diff/ > > > Testing > ------- > > None. > > > Thanks, > > Vinod Kone > >
