mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Neil Conway <neil.con...@gmail.com>
Subject Re: Review Request 50707: Added test for race between health checks and agent unregistration.
Date Tue, 13 Sep 2016 12:59:02 GMT


> On Sept. 12, 2016, 9:27 p.m., Vinod Kone wrote:
> > src/tests/slave_tests.cpp, lines 2869-2872
> > <https://reviews.apache.org/r/50707/diff/7/?file=1493683#file1493683line2869>
> >
> >     It's not clear to me what we are checking to verify that master behaves correctly
here? What would happen if master behaves incorrectly? Maybe add some comments?
> 
> Neil Conway wrote:
>     Well, we're checking the master does not segfault or hit a `CHECK` failure, which
it was doing before I fixed that bug and added this test :)
>     
>     Once `MockRegistrar` is introduced, we can validate that the master doesn't do any
registry writes in response to the `markUnreachable` dispatch. I'll add that as a separate
review at the end of the chain (we can't do it in this RR because `MockRegistrar` is added
in a later RR in this chain).
>     
>     I also improved the comments.

https://reviews.apache.org/r/51845/


- Neil


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/50707/#review148592
-----------------------------------------------------------


On Sept. 13, 2016, 12:49 p.m., Neil Conway wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/50707/
> -----------------------------------------------------------
> 
> (Updated Sept. 13, 2016, 12:49 p.m.)
> 
> 
> Review request for mesos and Vinod Kone.
> 
> 
> Bugs: MESOS-4049
>     https://issues.apache.org/jira/browse/MESOS-4049
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> This test tries to simulate a race between marking an agent unhealthy
> and receiving an `UnregisterSlaveMessage` for that agent.
> 
> Unfortunately, this test is a little fragile (we need to manually
> dispatch an event to the master process to simulate the action that
> would be taken by the slave observer).
> 
> 
> Diffs
> -----
> 
>   src/tests/slave_tests.cpp 2f3fa5fae634b6250f3c00bbef3077493f79af95 
> 
> Diff: https://reviews.apache.org/r/50707/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Neil Conway
> 
>


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