Thanks for the clarification. At the very least we should add a note to scheduler.hpp reflecting the current state.
On Thu, Feb 20, 2014 at 3:20 PM, Benjamin Hindman <b...@eecs.berkeley.edu>wrote: > The 'executorLost' was added for our first Apache release because we > thought we might want to implement it and we wanted it to be in the API. > Unfortunately, it never got implemented. > > Note, however, that from an API perspective both 'slaveLost' and > 'executorLost' are "best effort" and we should be clear about documenting > as much if/when the get implemented. > > > On Thu, Feb 20, 2014 at 10:49 AM, Vinod Kone <vinodk...@gmail.com> wrote: > > > I think this was just overlooked. Feel free to file a bug and send a > patch! > > > > > > On Wed, Feb 19, 2014 at 9:33 PM, Brenden Matthews < > > brenden.matth...@airbedandbreakfast.com> wrote: > > > > > Hey folks, > > > > > > I was trying to use the Scheduler::executorLost method[1], but to my > > dismay > > > I found that the callback was never called. > > > > > > A quick look at the source shows it's not ever referenced in the code, > > > except for a comment mentioning that it perhaps should be[2]. > > > > > > Any thoughts on this? Is the callback deprecated? It seems pretty > > straight > > > forward to implement. > > > > > > Best, > > > > > > Brenden > > > > > > [1]: > > > > > > > > > https://github.com/brndnmtthws/storm-mesos/blob/52dc535d5bcffb305ec008e30c70e3078dea957f/src/jvm/storm/mesos/MesosNimbus.java#L163-L168 > > > [2]: > > > > > > > > > https://github.com/apache/mesos/blob/master/src/master/master.cpp#L2126-L2128 > > > > > >
