> On April 5, 2013, 7:03 p.m., Ben Mahler wrote: > > src/slave/slave.cpp, line 1650 > > <https://reviews.apache.org/r/10278/diff/1/?file=277810#file277810line1650> > > > > You're calling get() without the Option having a value? This looks like > > it will fail the assertion in Option::get.
We set it right above this statement! > On April 5, 2013, 7:03 p.m., Ben Mahler wrote: > > src/slave/slave.cpp, line 1698 > > <https://reviews.apache.org/r/10278/diff/1/?file=277810#file277810line1698> > > > > What is the TODO to fix this "hack"? > > > > I'm missing the context so I don't see why we have to do this change? > > Can you elaborate in the review description? This was a long standing bug. As the comment states, if we don't have any pending tasks for a terminated executor, we were incorrectly assuming the executor as non-command executor. The proper way to fix this is to checkpoint this information, so that the slave can always know about an executor's type (command executor or not) even after a slave restart. But, we didn't want to checkpoint/recover yet another file. So, implemented a hacky solution to deduce it from the command. - Vinod ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/10278/#review18725 ----------------------------------------------------------- On April 4, 2013, 1:20 a.m., Vinod Kone wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/10278/ > ----------------------------------------------------------- > > (Updated April 4, 2013, 1:20 a.m.) > > > Review request for mesos, Benjamin Hindman and Ben Mahler. > > > Description > ------- > > See summary. > > > Diffs > ----- > > src/slave/slave.cpp 325231458a6883019436e7cc5a37f85f0f5735fa > > Diff: https://reviews.apache.org/r/10278/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Vinod Kone > >
