-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/71487/#review218028
-----------------------------------------------------------


Ship it!




The previous patch on recoverResources tried to avoid 
`allocation->offered->free` state transitions in two separate calls. The 
approach also suffers similar issues in that resources state update are splited 
into two calls when launching tasks: recover and transitionOfferedToAllocated. 
I think ideally, if we opt for this state transition approach, we should 
simplify all master intiated resources transitions into a single call where 
resources' state can be done atomocally for a single logical operation 
(e.g.launch a task on an offer, decline an offer).

But I think that will be a bigger refactor that we don't want to take right 
now. But can you add a TODO for the ideal solution here?

- Meng Zhu


On Sept. 27, 2019, 6:05 a.m., Andrei Sekretenko wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/71487/
> -----------------------------------------------------------
> 
> (Updated Sept. 27, 2019, 6:05 a.m.)
> 
> 
> Review request for mesos, Benjamin Mahler and Meng Zhu.
> 
> 
> Bugs: MESOS-9949
>     https://issues.apache.org/jira/browse/MESOS-9949
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> This is a prerequisite for tracking quota consumption in allocator.
> 
> 
> Diffs
> -----
> 
>   include/mesos/allocator/allocator.hpp 
> 2bab53ab5fb25931a724c20a039e1301983ba574 
>   src/master/allocator/mesos/allocator.hpp 
> 6921581745bf876ee42bcfb62b59245f23fcbf47 
>   src/master/allocator/mesos/hierarchical.hpp 
> d42124f534536ad044bff9878d242892fa042f24 
>   src/master/allocator/mesos/hierarchical.cpp 
> 47281542792486854b71c7e75dfa8ea3ee3ef3fc 
>   src/master/master.cpp 65994aa72ec4bfefdb09b88a62db910727bb0897 
>   src/tests/allocator.hpp 01e6d2ce3d0655ad408f605c7be84cd7c0f0d479 
> 
> 
> Diff: https://reviews.apache.org/r/71487/diff/3/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Andrei Sekretenko
> 
>

Reply via email to