mesos-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jie Yu" <yujie....@gmail.com>
Subject Re: Review Request 25858: Allowed co-mounted cgroup subsystems to enable Mesos on machines with systemd.
Date Mon, 22 Sep 2014 16:51:00 GMT


> On Sept. 22, 2014, 4:31 p.m., Timothy St. Clair wrote:
> > src/linux/cgroups.cpp, line 1795
> > <https://reviews.apache.org/r/25858/diff/1/?file=698357#file698357line1795>
> >
> >     I think I've missed something subtle, how did you bypass the cleanup semantics?

The isolators won't try to unmount and rm the hierarchies (i.e., it won't call cgroups::cleanup).

The only place we call cgroups::cleanup is in tests, I will do a sweep today to make sure
we don't accidentally umount systemd managed hierarchies.


> On Sept. 22, 2014, 4:31 p.m., Timothy St. Clair wrote:
> > src/slave/containerizer/isolators/cgroups/cpushare.cpp, line 112
> > <https://reviews.apache.org/r/25858/diff/1/?file=698359#file698359line112>
> >
> >     So does this rely on realpath squashing the symbolic link such that they are
= ?

In fact, we obtain the hierarchy from /proc/mounts. Yes, we also call realpath. So I think
this part is safe.


- Jie


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25858/#review54125
-----------------------------------------------------------


On Sept. 20, 2014, 12:21 a.m., Jie Yu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25858/
> -----------------------------------------------------------
> 
> (Updated Sept. 20, 2014, 12:21 a.m.)
> 
> 
> Review request for mesos, Ben Mahler, Ian Downes, Timothy St. Clair, and Vinod Kone.
> 
> 
> Bugs: MESOS-1195
>     https://issues.apache.org/jira/browse/MESOS-1195
> 
> 
> Repository: mesos-git
> 
> 
> Description
> -------
> 
> A dynamic version after discussed with Tim.
> https://reviews.apache.org/r/25695
> 
> Did a few consistency fixes as well.
> 
> 
> Diffs
> -----
> 
>   src/linux/cgroups.cpp 5093b4ca1ac17238234d96613b7f4ceab4373c48 
>   src/slave/containerizer/isolators/cgroups/cpushare.hpp d4df5f37e8d2e356d35ca40d799197a47393fa9a

>   src/slave/containerizer/isolators/cgroups/cpushare.cpp b1cad472a561e81422f980182fd24eb95701140a

>   src/slave/containerizer/isolators/cgroups/mem.cpp fb3db88af7b2ffa79272743f571c4c021c619c48

>   src/slave/containerizer/isolators/cgroups/perf_event.cpp ff047d37c1b2e659b18b5d4a1e97301192d05e55

>   src/slave/slave.cpp 28eb02852ddcc10efe589a8069dba9c895bc160e 
> 
> Diff: https://reviews.apache.org/r/25858/diff/
> 
> 
> Testing
> -------
> 
> make check
> sudo make check
> 
> 
> Thanks,
> 
> Jie Yu
> 
>


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