mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Guangya Liu <gyliu...@gmail.com>
Subject Re: Review Request 43884: Added allocator metrics for used quotas.
Date Wed, 24 Feb 2016 14:52:37 GMT


> On 二月 24, 2016, 1:26 p.m., Guangya Liu wrote:
> > src/master/allocator/mesos/hierarchical.cpp, lines 1077-1080
> > <https://reviews.apache.org/r/43884/diff/4/?file=1267201#file1267201line1077>
> >
> >     I think that we cannot say that the quota here is satisfied, but those are just
quotas that being used, some quotas may still starve.
> >     
> >     The following logic here https://github.com/apache/mesos/blob/master/src/master/allocator/mesos/hierarchical.cpp#L1230
is checking if a quota is satisified or not.
> >     
> >     Also for the comments in line 1079, what about use following which seems more
accurate because the quota also support allocating reserved resource for now. (Though non
revocable resources also include reserved resources, but I think that identifying the reserved
resources here may be more clear.)
> >     
> >     `Quota is satisfied with non-revocable and reserved resources.`
> 
> Benjamin Bannier wrote:
>     OK, I can see how someone might want to interpret *satisfied* as a binary, switched
to *used* instead.
>     
>     Re:l.1079, the comment reflects the current assumption that a quota is always satisfied
with non-revocable resources. That assumption is then used below when the metric value is
calculated. The comment is intended to just explain the assumptions made *here*.

I'm also thinking why need this comments here? Seems the comments does not related to the
logic here, comments?


- Guangya


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/43884/#review120496
-----------------------------------------------------------


On 二月 24, 2016, 1:49 p.m., Benjamin Bannier wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/43884/
> -----------------------------------------------------------
> 
> (Updated 二月 24, 2016, 1:49 p.m.)
> 
> 
> Review request for mesos, Alexander Rukletsov and Ben Mahler.
> 
> 
> Bugs: MESOS-4723
>     https://issues.apache.org/jira/browse/MESOS-4723
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Added allocator metrics for used quotas.
> 
> 
> Diffs
> -----
> 
>   src/master/allocator/mesos/hierarchical.hpp 3043888630b066505410d3b32c5b3f813cc458c1

>   src/master/allocator/mesos/hierarchical.cpp 5ef29f26ec8071f79c2f4f78dbe2bb0a613cc92d

>   src/tests/hierarchical_allocator_tests.cpp 5f771f02db9bd098f3cd36730cd84bf2f5e87a33

> 
> Diff: https://reviews.apache.org/r/43884/diff/
> 
> 
> Testing
> -------
> 
> make check (OS X)
> 
> I confirmed that this does not lead to general performance regressions in the allocator;
this is partially expected since the added code only inserts metrics in the allocator while
the actual work is perform asynchronously. These tests where performed with `HierarchicalAllocator_BENCHMARK_Test.DeclineOffers`
on an optimized build under OS X using clang(trunk) as compiler.
> 
> 
> Thanks,
> 
> Benjamin Bannier
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message