Re: Review Request 45051: Fixed flakiness in ContainerLoggerTest.LOGROTATE_RotateInSandbox.

2016-03-29 Thread Vinod Kone

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45051/#review125942
---


Ship it!




Ship It!

- Vinod Kone


On March 29, 2016, 7:32 p.m., Joseph Wu wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/45051/
> ---
> 
> (Updated March 29, 2016, 7:32 p.m.)
> 
> 
> Review request for mesos, Artem Harutyunyan and Vinod Kone.
> 
> 
> Bugs: MESOS-4961
> https://issues.apache.org/jira/browse/MESOS-4961
> 
> 
> Repository: mesos
> 
> 
> Description
> ---
> 
> This removes a race between the test and the global `ReapProcess`, which were 
> originally both calling `waitpid` on the logger subprocesses.  The test now 
> defers the reaping to the `ReapProcess` and just waits for all logger 
> subprocesses to exit naturally via `os::exists(pid)`.
> 
> Since the `ReapProcess` reaps on an interval, a `Clock::advance` was inserted 
> to skip the reap delay.
> 
> 
> Diffs
> -
> 
>   src/tests/container_logger_tests.cpp 
> 71101c31cee6a400b89cf285cf0a105d2d1534a8 
> 
> Diff: https://reviews.apache.org/r/45051/diff/
> 
> 
> Testing
> ---
> 
> make check
> 
> bin/mesos-tests.sh 
> --gtest_filter="ContainerLoggerTest.LOGROTATE_RotateInSandbox" 
> --gtest_repeat=1000 --gtest_break_on_failure
> 
> 
> Thanks,
> 
> Joseph Wu
> 
>



Re: Review Request 45051: Fixed flakiness in ContainerLoggerTest.LOGROTATE_RotateInSandbox.

2016-03-29 Thread Joseph Wu

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/45051/
---

(Updated March 29, 2016, 12:32 p.m.)


Review request for mesos, Artem Harutyunyan and Vinod Kone.


Changes
---

Changed this to a proper fix.


Bugs: MESOS-4961
https://issues.apache.org/jira/browse/MESOS-4961


Repository: mesos


Description (updated)
---

This removes a race between the test and the global `ReapProcess`, which were 
originally both calling `waitpid` on the logger subprocesses.  The test now 
defers the reaping to the `ReapProcess` and just waits for all logger 
subprocesses to exit naturally via `os::exists(pid)`.

Since the `ReapProcess` reaps on an interval, a `Clock::advance` was inserted 
to skip the reap delay.


Diffs (updated)
-

  src/tests/container_logger_tests.cpp 71101c31cee6a400b89cf285cf0a105d2d1534a8 

Diff: https://reviews.apache.org/r/45051/diff/


Testing (updated)
---

make check

bin/mesos-tests.sh 
--gtest_filter="ContainerLoggerTest.LOGROTATE_RotateInSandbox" 
--gtest_repeat=1000 --gtest_break_on_failure


Thanks,

Joseph Wu