> On Oct. 12, 2016, 10:56 p.m., Vinod Kone wrote:
> > src/executor/executor.cpp, line 734
> > <https://reviews.apache.org/r/52756/diff/1/?file=1531441#file1531441line734>
> >
> >     I think this should call `_shutdown()` to ensure ShutdownProcess is 
> > spawned  whenever we want to shutdown.

The actual plumbing happens through the `receive()` handler that first invokes 
the callback on the executor and then invokes `_shutdown()`. All other places 
in the code use `shutdown()` to inject the `SHUTDOWN` event.


- Anand


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/52756/#review152414
-----------------------------------------------------------


On Oct. 12, 2016, 12:15 a.m., Anand Mazumdar wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/52756/
> -----------------------------------------------------------
> 
> (Updated Oct. 12, 2016, 12:15 a.m.)
> 
> 
> Review request for mesos and Vinod Kone.
> 
> 
> Bugs: MESOS-6370
>     https://issues.apache.org/jira/browse/MESOS-6370
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Previously, the executor library used to commit suicide after the
> recovery timeout without invoking the executor's shutdown callback.
> This behavior was not consistent with the executor driver.
> 
> 
> Diffs
> -----
> 
>   src/executor/executor.cpp a0e5d8382fa3886b747ae92507dce17e75c2d749 
> 
> Diff: https://reviews.apache.org/r/52756/diff/
> 
> 
> Testing
> -------
> 
> Manual testing (It's hard to test this directly owing to the executor library 
> immediately invoking the `connected()` callback since the libprocess instance 
> initialized in tests is always active!)
> 
> Filed MESOS-6373 to address this.
> 
> 
> Thanks,
> 
> Anand Mazumdar
> 
>

Reply via email to