> On Jan. 11, 2016, 11:33 a.m., Klaus Ma wrote:
> > Would you add test case for that? It seems `make check` will also pass 
> > without this patch :).
> 
> Guangya Liu wrote:
>     I think that we need to update oversubscription_tests.cpp to use 
> usageSlack() instead of revocable() to get usage slack resources which is 
> more accurate.

That's not enough; after updating to use usageSlack(), the case in this patch 
is still not covered :). I'm going to add some mix case (both `USAGE_SLACK` & 
`ALLOCATION_SLACK` are reported) to catch such kind of issue.


- Klaus


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/42123/#review113670
-----------------------------------------------------------


On Jan. 11, 2016, 9:56 a.m., Guangya Liu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/42123/
> -----------------------------------------------------------
> 
> (Updated Jan. 11, 2016, 9:56 a.m.)
> 
> 
> Review request for mesos, Ben Mahler, Artem Harutyunyan, Joris Van 
> Remoortere, Joseph Wu, Klaus Ma, and Jian Qiu.
> 
> 
> Bugs: MESOS-4322
>     https://issues.apache.org/jira/browse/MESOS-4322
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> The current load qos controller is calculating executor used
> resources via revocable() API, but this is not right as the
> revocable() will include both USAGE SLACK and ALLOCATION SLACK.
> 
> This patch is updating the load qos controller only get USAGE
> SLACK revocable resources for executors.
> 
> The unit test was already covered in oversubscription_tests.cpp.
> 
> 
> Diffs
> -----
> 
>   src/slave/qos_controllers/load.cpp 52520d6220784423c09001b8bd59e090ebf787ec 
> 
> Diff: https://reviews.apache.org/r/42123/diff/
> 
> 
> Testing
> -------
> 
> make
> make check
> 
> 
> Thanks,
> 
> Guangya Liu
> 
>

Reply via email to