-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/52321/#review150658
-----------------------------------------------------------




src/slave/containerizer/composing.cpp (lines 491 - 497)
<https://reviews.apache.org/r/52321/#comment218692>

    Whoops, I didn't quite get this logic right when I introduced it in 
`_launch()`.
    
    If `launched` is true, we shouldn't return a failure here and we shouldn't 
set the `destroyed` Promise since we don't know if the container has been 
destroyed. It's only when `launched` is false that we can tell the caller we 
were able to "destroy" the container.
    
    However, in the nested container case, since we're not looping through 
containerizers, we could have destroy fail when launch returns false (but that 
probably just adds some complication here). =/


- Benjamin Mahler


On Sept. 27, 2016, 10:25 p.m., Vinod Kone wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/52321/
> -----------------------------------------------------------
> 
> (Updated Sept. 27, 2016, 10:25 p.m.)
> 
> 
> Review request for mesos, Anand Mazumdar, Benjamin Mahler, Gilbert Song, and 
> Jie Yu.
> 
> 
> Bugs: MESOS-6260
>     https://issues.apache.org/jira/browse/MESOS-6260
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Made composing containerizer nesting aware so that operators
> can enable both mesos and docker containerizers on agents.
> Ofcourse docker containerizer is not nesting aware.
> 
> 
> Diffs
> -----
> 
>   src/slave/containerizer/composing.hpp 
> 9be780201ec2efae02d560fa3560fe12f1d53a50 
>   src/slave/containerizer/composing.cpp 
> 179304a9f0e1e4ef7e2946feea97c0897e24f5ab 
>   src/tests/default_executor_tests.cpp 
> 3d6cbd92b916f38325f9f23245cc602bae4f83e3 
> 
> Diff: https://reviews.apache.org/r/52321/diff/
> 
> 
> Testing
> -------
> 
> sudo make check
> 
> 
> Thanks,
> 
> Vinod Kone
> 
>

Reply via email to