mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Vinod Kone <vinodk...@gmail.com>
Subject Re: Review Request 44275: Fixed flakiness in tests using the scheduler library.
Date Thu, 03 Mar 2016 01:10:21 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/44275/#review121754
-----------------------------------------------------------




src/tests/mesos.hpp (lines 992 - 994)
<https://reviews.apache.org/r/44275/#comment183564>

    MESOS-4029 is talking about a bunch of issues. The final thing it says is to use a shared_ptr.
There is nothing about why there needs to be a stop() and this settle().
    
    Can you explain here with a little more detail for posterity? both the stop() and settle().


- Vinod Kone


On March 2, 2016, 9:08 p.m., Anand Mazumdar wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/44275/
> -----------------------------------------------------------
> 
> (Updated March 2, 2016, 9:08 p.m.)
> 
> 
> Review request for mesos and Vinod Kone.
> 
> 
> Bugs: MESOS-4029
>     https://issues.apache.org/jira/browse/MESOS-4029
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> This change fixes the tests using the scheduler library by
> explicitly invoking `stop()` from the testing interface. This
> ensures that no further callbacks are delivered to the scheduler.
> 
> For one-off async callbacks that are already on the libprocess queue
> we need to do a `Clock::settle` to ensure they are executed before
> the mock object goes out of scope.
> 
> 
> Diffs
> -----
> 
>   src/tests/mesos.hpp 1d4f075e470a60040e17b9f011aea6202310c437 
> 
> Diff: https://reviews.apache.org/r/44275/diff/
> 
> 
> Testing
> -------
> 
> make check + Induced a sleep in the async callbacks. Previously, the test used to crash
but with the fix it no longer crashes.
> 
> 
> Thanks,
> 
> Anand Mazumdar
> 
>


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