> On Aug. 14, 2016, 2:38 p.m., Qian Zhang wrote:
> > include/mesos/mesos.proto, lines 1381-1383
> > <https://reviews.apache.org/r/51061/diff/1/?file=1472355#file1472355line1381>
> >
> >     Do we allow a custom executor to launch multiple task groups and each 
> > task group is in its own network? If so, I think we may need to add 
> > `NetworkInfo` in this message.

That's an interesting idea. We haven't really discussed separate nework 
namespaces for different task groups all owned by a single, custom executor. 
What's the use case here? You want multiple task groups that may share access 
to the same storage but each group has a distinct netns?


- James


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/51061/#review145723
-----------------------------------------------------------


On Aug. 12, 2016, 11:12 p.m., Vinod Kone wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/51061/
> -----------------------------------------------------------
> 
> (Updated Aug. 12, 2016, 11:12 p.m.)
> 
> 
> Review request for mesos, Anand Mazumdar, Benjamin Hindman, Benjamin Mahler, 
> James DeFelice, and Jie Yu.
> 
> 
> Bugs: MESOS-6036
>     https://issues.apache.org/jira/browse/MESOS-6036
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> A scheduler can use the LAUNCH_GROUP offer operation to send a
> `TaskGroup` to an executor.
> 
> 
> Diffs
> -----
> 
>   include/mesos/mesos.proto af29dab2af4c28e257ad518c68d3f1af7cad0d06 
>   include/mesos/v1/mesos.proto a836f6e792944726ec6f8f66f77f403894377d05 
>   src/common/resources.cpp 6b7af9179121efbdc5c29484eb042778bcea8288 
>   src/master/master.cpp 0bd1a3490a86fede86a3f5f62ce4745b65aae258 
>   src/v1/resources.cpp 03ee0cb0bb5abe7fc1ae4cb47c5b6dbcd8d11998 
> 
> Diff: https://reviews.apache.org/r/51061/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Vinod Kone
> 
>

Reply via email to