mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Benjamin Mahler <bmah...@apache.org>
Subject Re: Review Request 64303: Tracked resource reservations in the allocator.
Date Sat, 09 Dec 2017 00:18:12 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/64303/#review192962
-----------------------------------------------------------



Looks good! Just some minor suggestions below


src/master/allocator/mesos/hierarchical.hpp
Lines 453-457 (patched)
<https://reviews.apache.org/r/64303/#comment271372>

    Can you comment on the lifecycle of the entries in the map? My intuition would guess that
we only store entries when they have reservations, if a role doesn't have any reservations
we don't bother putting it in this map?
    
    i.e. there will never be an empty resources, in that case the key will be absent?



src/master/allocator/mesos/hierarchical.hpp
Lines 546-555 (patched)
<https://reviews.apache.org/r/64303/#comment271373>

    Perhaps a TODO here that reflects your plan based on the ticket for fixing the fairness
issues when there are unallocated reservations?
    
    Would be helpful for others.



src/master/allocator/mesos/hierarchical.hpp
Lines 552-555 (patched)
<https://reviews.apache.org/r/64303/#comment271375>

    Also, as it stands it's a little unclear naming-wise to me that that `updateReservations`
has a relationship to the track and untrack functions.
    
    But maybe we can just simplify and remove this? In favor of just having the caller untrack
old and track the new.



src/master/allocator/mesos/hierarchical.cpp
Lines 2365-2366 (patched)
<https://reviews.apache.org/r/64303/#comment271838>

    Can you match the function order in the header? So:  track defined before untrack



src/master/allocator/mesos/hierarchical.cpp
Lines 2368-2369 (patched)
<https://reviews.apache.org/r/64303/#comment271839>

    s/name/role/ here and elsewhere



src/master/allocator/mesos/hierarchical.cpp
Lines 2374 (patched)
<https://reviews.apache.org/r/64303/#comment271840>

    how about scalarQuantitesToUntrack?
    
    s/&//



src/master/allocator/mesos/hierarchical.cpp
Lines 2374-2375 (patched)
<https://reviews.apache.org/r/64303/#comment271844>

    A note here that we need to call toUnreserved to strip the static reservation metadata
would be helpful for the reader.



src/master/allocator/mesos/hierarchical.cpp
Lines 2391 (patched)
<https://reviews.apache.org/r/64303/#comment271841>

    how about scalarQuantitesToTrack?
    
    s/&//



src/master/allocator/mesos/hierarchical.cpp
Lines 2391-2392 (patched)
<https://reviews.apache.org/r/64303/#comment271845>

    A note here that we need to call toUnreserved to strip the static reservation metadata
would be helpful for the reader.



src/master/allocator/mesos/hierarchical.cpp
Lines 2394-2398 (patched)
<https://reviews.apache.org/r/64303/#comment271842>

    This can just be:
    
    ```
    reservationScalarQuantities[name] += scalarQuantitesToTrack;
    ```


- Benjamin Mahler


On Dec. 4, 2017, 11:18 p.m., Meng Zhu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/64303/
> -----------------------------------------------------------
> 
> (Updated Dec. 4, 2017, 11:18 p.m.)
> 
> 
> Review request for mesos, Alexander Rukletsov, Benjamin Mahler, and Michael Park.
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Resource reservations need to be tracked to make
> sure quota limit will not be exceeded in the presence
> of resource reservation.
> 
> 
> Diffs
> -----
> 
>   src/master/allocator/mesos/hierarchical.hpp 41ffe17def133d7e735afa26a6fedd154e51f4b4

>   src/master/allocator/mesos/hierarchical.cpp 20b908cc891f9f9be3045cd9f8be83a11d37ab78

> 
> 
> Diff: https://reviews.apache.org/r/64303/diff/2/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Meng Zhu
> 
>


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