> On March 31, 2016, 2:12 a.m., Jie Yu wrote:
> > src/slave/containerizer/mesos/isolators/network/cni/cni.cpp, line 339
> > <https://reviews.apache.org/r/45383/diff/4/?file=1319540#file1319540line339>
> >
> >     I think this is not needed because if recover fails, slave will restart.

Agree, but I see other isolators do the similar, e.g., 
CgroupsCpushareIsolatorProcess::recover(), 
CgroupsMemIsolatorProcess::recover(), etc., maybe we need to fix them as well?


> On March 31, 2016, 2:12 a.m., Jie Yu wrote:
> > src/slave/containerizer/mesos/isolators/network/cni/cni.cpp, lines 887-890
> > <https://reviews.apache.org/r/45383/diff/4/?file=1319540#file1319540line887>
> >
> >     Or agent crashes right before removing the containerDir?

Yes! Let me add it in the comment.


> On March 31, 2016, 2:12 a.m., Jie Yu wrote:
> > src/slave/containerizer/mesos/isolators/network/cni/cni.cpp, lines 891-908
> > <https://reviews.apache.org/r/45383/diff/4/?file=1319540#file1319540line891>
> >
> >     This basically duplicates what's in the cleanup. I am wondering if we 
> > should make cleanup more robust to tolerate a container dir without network 
> > name dirs. In other words, remove the logic here, change cleanup so that it 
> > can handle empty network dirs case.

Yes, we can do it, but then we have to remove 
`CHECK(infos.contains(containerId))` from `_cleanup()` and make the logic in 
the `cleanup()` a little obscure. Anyway, I have updated the code accordingly, 
please take a look and let me know for any further comments :-)


> On March 31, 2016, 2:12 a.m., Jie Yu wrote:
> > src/slave/containerizer/mesos/isolators/network/cni/cni.cpp, line 864
> > <https://reviews.apache.org/r/45383/diff/4/?file=1319540#file1319540line864>
> >
> >     OK, now I realized that `getNetworkInfoDir` is a little confusing with 
> > `getNetworkDir`. Can we `s/getNetworkInfoDir/getContainerDir/` in a 
> > subsequent patch?

Sure, I have renamed it in the patch: https://reviews.apache.org/r/45532/


- Qian


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45383/#review126143
-----------------------------------------------------------


On March 30, 2016, 9:39 p.m., Qian Zhang wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/45383/
> -----------------------------------------------------------
> 
> (Updated March 30, 2016, 9:39 p.m.)
> 
> 
> Review request for mesos, Avinash sridharan, Gilbert Song, and Jie Yu.
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Implemented recover() method of "network/cni" isolator.
> 
> 
> Diffs
> -----
> 
>   src/slave/containerizer/mesos/isolators/network/cni/cni.hpp 
> 873e0c52475f4868e611bd24a6782ad5eb261a99 
>   src/slave/containerizer/mesos/isolators/network/cni/cni.cpp 
> 1c8e231813c0579b79681c5d18b1f799a727ead7 
>   src/slave/containerizer/mesos/isolators/network/cni/paths.hpp 
> f627ec9499a34ca104d2c1a4d28e1d2f4b849f64 
>   src/slave/containerizer/mesos/isolators/network/cni/paths.cpp 
> 611f3869402b9033081b7f9ecc1bdf006f61918b 
>   src/slave/containerizer/mesos/isolators/network/cni/spec.hpp 
> 6a3c33645bab73edaf5af4d298a671852ea59c46 
>   src/slave/containerizer/mesos/isolators/network/cni/spec.cpp 
> 5b5f904def9ef6dcc4462a03a2d024ad4eb3d946 
> 
> Diff: https://reviews.apache.org/r/45383/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Qian Zhang
> 
>

Reply via email to