mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Marco Massenzio" <ma...@mesosphere.io>
Subject Re: Review Request 37532: Add QUIESCE call interface to the scheduler
Date Thu, 20 Aug 2015 19:39:35 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37532/#review95997
-----------------------------------------------------------



include/mesos/scheduler.hpp (line 273)
<https://reviews.apache.org/r/37532/#comment151190>

    indent is off?



include/mesos/scheduler.hpp (line 424)
<https://reviews.apache.org/r/37532/#comment151195>

    indent



include/mesos/scheduler/scheduler.proto (line 173)
<https://reviews.apache.org/r/37532/#comment151191>

    micro-nit: the // is not aligned with lines above.



include/mesos/scheduler/scheduler.proto (lines 314 - 315)
<https://reviews.apache.org/r/37532/#comment151192>

    This comments does not read well: what is the timeout? also, it would be good to have
a bit of info about what the filters are
    
    (eg, are they 'inclusion' or 'exclusion' filters? etc.)



src/master/master.cpp (line 2507)
<https://reviews.apache.org/r/37532/#comment151193>

    IMO we should LOG(INFO) here to state this is not implemented yet or something.
    
    Otherwise, say people see this and start using it, then nothing happens, look at the logs
(which give the impression that the "Processing QUIESCE" was successfully called) and will
report it as a bug.



src/sched/sched.cpp (line 1953)
<https://reviews.apache.org/r/37532/#comment151194>

    indent off



src/tests/scheduler_tests.cpp (line 1003)
<https://reviews.apache.org/r/37532/#comment151196>

    so, this is a good test, but I would really like to see one where we ask Master to keep
quiet for a while and we don't get offers during that period of time, then we start getting
again.
    
    It may require some fiddling around with `Clock`s and all that jazz, but it would give
us more confidence that this whole thing works.
    
    Also - some tests around more complex filtering (if any? maybe this is there is, then
it's fine).


- Marco Massenzio


On Aug. 17, 2015, 1:47 p.m., Guangya Liu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37532/
> -----------------------------------------------------------
> 
> (Updated Aug. 17, 2015, 1:47 p.m.)
> 
> 
> Review request for mesos and Vinod Kone.
> 
> 
> Bugs: MESOS-3037
>     https://issues.apache.org/jira/browse/MESOS-3037
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> This is just part of MESOS-3037, this patch only add the interface
> of QUIESCE call.
> 
> 
> Diffs
> -----
> 
>   include/mesos/scheduler.hpp ee198b6955882f4f31466ca05429ca16fbf2f5cd 
>   include/mesos/scheduler/scheduler.proto 89daf8a6b74057ee156b3ad691397e76fcb835b8 
>   include/mesos/v1/scheduler/scheduler.proto bd5e82a614b1163b29f9b20e562208efa1ba4b55

>   src/master/master.hpp 0432842d77beba024c7895291ca410964bae96be 
>   src/master/master.cpp c5e6c6f3304060d4c92d52851951f10bc432500e 
>   src/sched/sched.cpp 012af0508eeceeccd168b29f36fa258d20b28c21 
>   src/tests/scheduler_tests.cpp 77c26353afc33f5099be2d1e597ffc630e559968 
> 
> Diff: https://reviews.apache.org/r/37532/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Guangya Liu
> 
>


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