> On Jan. 22, 2013, 4:21 a.m., Benjamin Hindman wrote:
> > src/tests/reaper_tests.cpp, lines 45-47
> > <https://reviews.apache.org/r/8570/diff/4/?file=250446#file250446line45>
> >
> >     In other tests, we used 'ASSERT_NE(-1, pipe(pipes));'. While I like the 
> > idea of getting more debug info from the PLOG, I also like the idea of only 
> > aborting this test case. Does a PASSERT exist?

agreed on aborting only this test case. couldn't find anything akin to PASSERT.


> On Jan. 22, 2013, 4:21 a.m., Benjamin Hindman wrote:
> > src/tests/reaper_tests.cpp, line 54
> > <https://reviews.apache.org/r/8570/diff/4/?file=250446#file250446line54>
> >
> >     I'm guilty of this too, but we really should not treat 'pid' as a 
> > boolean, because it is not. We should probably do a general "sweep" through 
> > the code and rectify everywhere.

done. here and everywhere else.


> On Jan. 22, 2013, 4:21 a.m., Benjamin Hindman wrote:
> > src/tests/reaper_tests.cpp, line 70
> > <https://reviews.apache.org/r/8570/diff/4/?file=250446#file250446line70>
> >
> >     This is a better place for doing something like PLOG, but you'll want 
> > to use perror since a child process might have the dreaded glog fork 
> > deadlock. Note that you _DO_ want to kill the entire process in this case 
> > otherwise you'll have two processes executing test cases, which gets very 
> > confusing for the person running tests.

agreed. fixed


> On Jan. 22, 2013, 4:21 a.m., Benjamin Hindman wrote:
> > src/tests/reaper_tests.cpp, lines 79-80
> > <https://reviews.apache.org/r/8570/diff/4/?file=250446#file250446line79>
> >
> >     If you want to be sure, use a pipe.

I decided to be lazy and use a while loop instead.


> On Jan. 22, 2013, 4:21 a.m., Benjamin Hindman wrote:
> > src/slave/reaper.cpp, line 47
> > <https://reviews.apache.org/r/8570/diff/4/?file=250444#file250444line47>
> >
> >     Cool. I like the cleanup, but I still think the long term solution here 
> > is to return a Future that gets satisfied once a pid is gone (reaped by us 
> > or otherwise).

as discussed offline, I'm going to make this a TODO for now.


- Vinod


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/8570/#review15556
-----------------------------------------------------------


On Jan. 19, 2013, 1:52 a.m., Vinod Kone wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/8570/
> -----------------------------------------------------------
> 
> (Updated Jan. 19, 2013, 1:52 a.m.)
> 
> 
> Review request for mesos, Benjamin Hindman and Ben Mahler.
> 
> 
> Description
> -------
> 
> Needed this to properly monitor the exit status of re-connected executors, as 
> they will be parented by INIT. 
> 
> 
> Diffs
> -----
> 
>   src/Makefile.am 0ab59b75b2955c532d0833f132bdaffe323838d0 
>   src/linux/proc.hpp e0825a4a9f9e2763e0c25d7319f220bfe7c7c29c 
>   src/linux/proc.cpp 8a0fc48dc9769df35d682ece477246b2df2fc0d4 
>   src/slave/cgroups_isolation_module.cpp 
> 63cefc33cf34eebb82db5d8448b751be8652fa36 
>   src/slave/lxc_isolation_module.cpp 30cff2a49339bb07030727d30352536a0a22d58c 
>   src/slave/process_based_isolation_module.cpp 
> 3d50a4b652e4e09dd57e744e408c8fb79ff3fbf5 
>   src/slave/reaper.hpp b9aa62daa42bdaa736ade43884982529ba3d4bb1 
>   src/slave/reaper.cpp c0ee4b4c07fd792bcb39455b666808b712eb32c2 
>   src/tests/master_tests.cpp 948ab5dff34eeba1f3ce593a864ddf282c8b19ed 
>   src/tests/reaper_tests.cpp PRE-CREATION 
>   src/tests/utils.hpp be457117515ee727af101370b26bf9188afb8f45 
>   third_party/libprocess/Makefile.am 11f8b36d19313f65f1b191a781c6d5ed90130099 
>   third_party/libprocess/include/stout/proc.hpp PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/8570/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Vinod Kone
> 
>

Reply via email to