-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/63523/#review190160
-----------------------------------------------------------


Fix it, then Ship it!





src/master/allocator/mesos/hierarchical.cpp
Line 828 (original), 768 (patched)
<https://reviews.apache.org/r/63523/#comment267424>

    This can be a `const` ref.



src/master/allocator/mesos/hierarchical.cpp
Lines 888-896 (original), 809-817 (patched)
<https://reviews.apache.org/r/63523/#comment267425>

    This could be more general, and I am not sure I agree with the general 
direction the `TODO` outlines. In the end we only need to handle conversions 
which actually consume resources here, and can drop all which don't. This can 
in principle be independent from e.g., launches on shared resources.



src/master/master.cpp
Lines 4890 (patched)
<https://reviews.apache.org/r/63523/#comment267419>

    Not yours, but this could be a `const` ref.



src/master/master.cpp
Lines 4888-4889 (original)
<https://reviews.apache.org/r/63523/#comment267422>

    This is weird. Any idea why we did inject this information into allocators? 
Is there any expectation that a custom allocator should be able to access this 
information?



src/tests/hierarchical_allocator_tests.cpp
Lines 5001-5003 (original), 5015-5017 (patched)
<https://reviews.apache.org/r/63523/#comment267426>

    This is weird now since `launch` is only used to calculate the expectation 
below.
    
    We should either try to inject it in the allocator just like `create`, or 
try to get rid of it completely.


- Benjamin Bannier


On Nov. 2, 2017, 8:36 p.m., Jie Yu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/63523/
> -----------------------------------------------------------
> 
> (Updated Nov. 2, 2017, 8:36 p.m.)
> 
> 
> Review request for mesos, Benjamin Bannier, Benjamin Mahler, Gaston Kleiman, 
> Greg Mann, Jan Schlicht, Vinod Kone, and Jiang Yan Xu.
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Instead of taking offer operations, it's more clear to make allocator
> not aware of offer operations. This patch changes the interface of the
> `updateAllocation` call to take `ResourceConversion`s.
> 
> 
> Diffs
> -----
> 
>   include/mesos/allocator/allocator.hpp 
> 537658bf4e68ac7f271248af0e1e35df2b19aef8 
>   src/master/allocator/mesos/allocator.hpp 
> 903edf68e812e022cddedd40a2cb51726b2a181b 
>   src/master/allocator/mesos/hierarchical.hpp 
> c2346054b2c98516f15ab8ce2dc798224ff4def4 
>   src/master/allocator/mesos/hierarchical.cpp 
> 848e9da1b61f411449c0cd2409e98ec733ac10ee 
>   src/master/master.cpp e047462018af50b9047b8939e922d80c4771fb28 
>   src/tests/allocator.hpp f1c0d148886b7b2936918b7d7c77c03ee766ee4a 
>   src/tests/hierarchical_allocator_tests.cpp 
> 48b48ad70f44cc2232c2a29699267027f9937b8a 
> 
> 
> Diff: https://reviews.apache.org/r/63523/diff/1/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Jie Yu
> 
>

Reply via email to