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 45370: Implemented prepare() for volume isolator.
Date Thu, 21 Apr 2016 23:33:01 GMT


> On 四月 21, 2016, 9:36 p.m., Gilbert Song wrote:
> > src/slave/containerizer/mesos/isolators/docker/volume/isolator.hpp, line 77
> > <https://reviews.apache.org/r/45370/diff/8/?file=1355000#file1355000line77>
> >
> >     Should we use a hashset for `dockerVolumeInfos`?
> >     
> >     I am thinking if a user has two `DockerVolumleInfo` with exactly the same `dockerVolume`,
`containerPath` and `driverOptions`, do we still bind mount them twice?

+1, I think that we should throw error for such case, what do you think?


> On 四月 21, 2016, 9:36 p.m., Gilbert Song wrote:
> > src/slave/containerizer/mesos/isolators/docker/volume/isolator.cpp, lines 259-274
> > <https://reviews.apache.org/r/45370/diff/8/?file=1355001#file1355001line259>
> >
> >     I would prefer remove `vector<string> messages` and return a failure once
you get the first non-ready future. Please add necessary failure message.

Gilbert, I was following similar logic in https://github.com/apache/mesos/blob/master/src/slave/containerizer/mesos/isolators/network/cni/cni.cpp#L671-L680
, I think that this can give end user a more clear message for all errors and is helpful to
trouble shooting, what do you think?


> On 四月 21, 2016, 9:36 p.m., Gilbert Song wrote:
> > src/slave/containerizer/mesos/isolators/docker/volume/isolator.cpp, line 213
> > <https://reviews.apache.org/r/45370/diff/8/?file=1355001#file1355001line213>
> >
> >     const string volumePath

Why not `string&`? Any more comments? ;-)


> On 四月 21, 2016, 9:36 p.m., Gilbert Song wrote:
> > src/slave/containerizer/mesos/isolators/docker/volume/isolator.cpp, lines 141-144
> > <https://reviews.apache.org/r/45370/diff/8/?file=1355001#file1355001line141>
> >
> >     Not yours, just leave a note here.
> >     
> >     @jieyu, currently we have different places (e.g., slave.cpp, filesystem/linux.cpp,
mesos/containerizer.cpp) explicitly exclude diff volume type.
> >     
> >     As we may have more volume types in the future, we should find a universal way
to handle diff volume type exclussion.

I will update here as following as we now have only one source: DOCKER_VOLUME.

if (!volume.has_source()) {
  continue;
}


- Guangya


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45370/#review129959
-----------------------------------------------------------


On 四月 21, 2016, 2:21 p.m., Guangya Liu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/45370/
> -----------------------------------------------------------
> 
> (Updated 四月 21, 2016, 2:21 p.m.)
> 
> 
> Review request for mesos, David vonThenen, Gilbert Song, haosdent huang, and Jie Yu.
> 
> 
> Bugs: MESOS-5013
>     https://issues.apache.org/jira/browse/MESOS-5013
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Implemented prepare() for volume isolator.
> 
> 
> Diffs
> -----
> 
>   src/slave/containerizer/mesos/isolators/docker/volume/isolator.hpp bedc687cc280d0b721fb84801039fd3614364cca

>   src/slave/containerizer/mesos/isolators/docker/volume/isolator.cpp 915e5ae755a55a02b7dfcda88165f27346cad955

>   src/slave/slave.cpp d82dec2b10d496065013eb4ad6a35dc054b72553 
> 
> Diff: https://reviews.apache.org/r/45370/diff/
> 
> 
> Testing
> -------
> 
> make
> make check
> 
> Will update `Sequence` later.
> 
> 
> Thanks,
> 
> Guangya Liu
> 
>


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