> On Nov. 17, 2015, 7:27 p.m., Vinod Kone wrote:
> > src/tests/containerizer/cgroups_tests.cpp, lines 560-568
> > <https://reviews.apache.org/r/38234/diff/1/?file=1066658#file1066658line560>
> >
> >     instead of asserting, it would be better if we can disable the test 
> > automatically if we detect that swap is enabled.
> >     
> >     for example, we can change the test name to ROOT_CGROUPS_OOM_Listen
> >     
> >     and in environment.cpp, disable the test if 'OOM' token is present in 
> > the test name and swap is enabled.
> >     
> >     
> >     also, does the swap issue not affect the balloon framewok test?

Yeah, I wanted to do that to disable other tests dymaically before and final 
result was just to use this style, so I did the same to be consistent.

re-read the code again for balloon frame tests: slave has 96MB memory, executor 
uses 64MB, 32 left to the task, balloon() allocates 64MB at a time, which 
basically means first allocation request will cause OOM (64 > 32).

If the step was set to say do 3 times 16MBs, (3x16 > 32), the extra 16MB will 
be eaten by swap and not cause a OOM if swap is enabled.


- Chi


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/38234/#review106919
-----------------------------------------------------------


On Sept. 9, 2015, 9:33 p.m., Chi Zhang wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/38234/
> -----------------------------------------------------------
> 
> (Updated Sept. 9, 2015, 9:33 p.m.)
> 
> 
> Review request for mesos and Jie Yu.
> 
> 
> Bugs: mesos-2918
>     https://issues.apache.org/jira/browse/mesos-2918
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Check if swap is enabled before running memory pressure related tests.
> 
> 
> Diffs
> -----
> 
>   src/tests/containerizer/cgroups_tests.cpp 
> 75a3bc0009c037dc18ce319db2eb44630f083e8c 
> 
> Diff: https://reviews.apache.org/r/38234/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Chi Zhang
> 
>

Reply via email to