> On March 8, 2016, 5:54 p.m., Guangya Liu wrote:
> > src/tests/master_maintenance_tests.cpp, lines 501-503
> > <https://reviews.apache.org/r/44537/diff/2/?file=1292383#file1292383line501>
> >
> >     @Joseph, just a question, how can this guarantee that there is no other 
> > events after the Clock::settle()?
> 
> Joseph Wu wrote:
>     This is actually a "best-effort" check rather than a guarantee.  It's 
> essentially a guard against sending 2 inverse offers right after another (the 
> bug you fixed in the previous patch).
> 
> Guangya Liu wrote:
>     Yes, I also add similar logic based on comments from @Anand 
> https://reviews.apache.org/r/44258/ and assert the event is pending for such 
> case, I do not know how would you rebase this patch or do you think is it 
> necessary to add such checking? I think that we need the checking proposed by 
> @Anand as this can make sure no other offers coming.
> 
> Anand Mazumdar wrote:
>     I hadn't seen this patch when I had raised the initial comment on r44258. 
> The eventual plan is to move all the maintenance tests to the new scheduler 
> mock interface for testing like this patch does. 
>     
>     I am wondering if it would be a good idea to make r44258 take care of the 
> actual implementation i.e. fixing the bug in master and r44537 takes care of 
> the testing details? 
>     
>     [~gyliu] How does this sound to you/Joseph?
> 
> Guangya Liu wrote:
>     Yes, that's also my think, I will remove the logic of `Clock::settle()` 
> but keep the logic limitted in master and leave the test update to this 
> patch. But I still want to know how we can check that there is no event 
> coming after `Clock::settle()` ;-)
> 
> Anand Mazumdar wrote:
>     I suggested to Joseph to advance the clock by `allocation_interval` and 
> then do a `Clock::settle` to _ensure_ that we do not get another offer to no 
> longer have it as "best-effort".

Added it.  I should emphasize that this check is still best-effort :)  The 
`clock::advance` only makes it a slightly "better" effort.


- Joseph


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/44537/#review122652
-----------------------------------------------------------


On March 10, 2016, 1:08 p.m., Joseph Wu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/44537/
> -----------------------------------------------------------
> 
> (Updated March 10, 2016, 1:08 p.m.)
> 
> 
> Review request for mesos, Anand Mazumdar, Artem Harutyunyan, and Joris Van 
> Remoortere.
> 
> 
> Bugs: MESOS-4831
>     https://issues.apache.org/jira/browse/MESOS-4831
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Based on the modified test Anand posted in the description of MESOS-4831.
> 
> 
> Diffs
> -----
> 
>   src/tests/master_maintenance_tests.cpp 
> 3faa8136cf57276295553910319480028f433e4c 
> 
> Diff: https://reviews.apache.org/r/44537/diff/
> 
> 
> Testing
> -------
> 
> make check OSX
> 
> 
> Thanks,
> 
> Joseph Wu
> 
>

Reply via email to