mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Avinash sridharan <avin...@mesosphere.io>
Subject Re: Review Request 42586: Defined the NetClsHandleManager class.
Date Wed, 03 Feb 2016 06:44:31 GMT


> On Feb. 1, 2016, 5:49 p.m., Jie Yu wrote:
> > src/slave/containerizer/mesos/isolators/cgroups/net_cls.hpp, lines 51-52
> > <https://reviews.apache.org/r/42586/diff/6/?file=1227492#file1227492line51>
> >
> >     `s/majHandle/_major/`
> >     `s/minHandle/_minor/`
> >     
> >     Also, remove the space right after `NetClsHandle`

Using primary and secondary instead of major and minor.


> On Feb. 1, 2016, 5:49 p.m., Jie Yu wrote:
> > src/slave/containerizer/mesos/isolators/cgroups/net_cls.hpp, lines 65-75
> > <https://reviews.apache.org/r/42586/diff/6/?file=1227492#file1227492line65>
> >
> >     Why do we need this?
> 
> Avinash sridharan wrote:
>     Was using this data structure to pass on the usage of handles for a given major handle.
It was useful in unit-testing, but thought will be useful in implementing a usage interface
to the NetClsHandleMgr class. Should I keep it?

Removed this structure. We no longer need it.


> On Feb. 1, 2016, 5:49 p.m., Jie Yu wrote:
> > src/slave/containerizer/mesos/isolators/cgroups/net_cls.hpp, lines 78-85
> > <https://reviews.apache.org/r/42586/diff/6/?file=1227492#file1227492line78>
> >
> >     I would suggest that we don't define a new data structure for this. Can you
just inline it in NetClsHandleManager?

Simplified the data structure. We no longer need an explicit NetClsHandles structure.


> On Feb. 1, 2016, 5:49 p.m., Jie Yu wrote:
> > src/slave/containerizer/mesos/isolators/cgroups/net_cls.hpp, lines 108-113
> > <https://reviews.apache.org/r/42586/diff/6/?file=1227492#file1227492line108>
> >
> >     We do you need this in this patch?
> 
> Avinash sridharan wrote:
>     There are quite a few places where we are throwing erros and need to convert the
major and minor handle into a hex string, hence thought implement this as a static function
(maybe inline) . Should I keep this?

Removed this structure from this patch. Will re-introduce it when we introduce the implementation.


> On Feb. 1, 2016, 5:49 p.m., Jie Yu wrote:
> > src/slave/containerizer/mesos/isolators/cgroups/net_cls.hpp, line 129
> > <https://reviews.apache.org/r/42586/diff/6/?file=1227492#file1227492line129>
> >
> >     What's the key for this map?

Added a comment to elaborate on the key.


- Avinash


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/42586/#review117225
-----------------------------------------------------------


On Feb. 3, 2016, 6:44 a.m., Avinash sridharan wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/42586/
> -----------------------------------------------------------
> 
> (Updated Feb. 3, 2016, 6:44 a.m.)
> 
> 
> Review request for mesos and Jie Yu.
> 
> 
> Bugs: MESOS-4345
>     https://issues.apache.org/jira/browse/MESOS-4345
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> This class will be responsible for keeping track of the free and allocated
> net_cls handles.
> 
> 
> Diffs
> -----
> 
>   src/slave/containerizer/mesos/isolators/cgroups/net_cls.hpp b4bc52114389d1c1efce2830f4292bd89bb0de7c

>   src/slave/containerizer/mesos/isolators/cgroups/net_cls.cpp ddc1bf0939e5e8995e6f34fe7b8509b51704f63e

> 
> Diff: https://reviews.apache.org/r/42586/diff/
> 
> 
> Testing
> -------
> 
> make
> 
> 
> Thanks,
> 
> Avinash sridharan
> 
>


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