> On Sept. 16, 2014, 11:20 a.m., Ben Mahler wrote:
> > src/master/detector.cpp, lines 396-404
> > <https://reviews.apache.org/r/25663/diff/1/?file=689921#file689921line396>
> >
> >     Any way to explicitly ignore the no node case?
> >     
> >     Here we're ignoring all failures, which is a bit scary, since it's not 
> > obvious why the more serious errors would be caught elsewhere.
> >     
> >     I haven't put a lot of thought into this, could we return a 
> > Future<Option<string> > data? Or should we consider wrapping our things in 
> > some kind of ZKOperation<string> which lets us look at the various error 
> > cases?
> >     
> >     Happy to chat further!

Yeah I was (unnecessarily) concerned about the ambiguity of returning a None if 
the method is changed to: "Result<Option<string> > GroupProcess::doData(const 
Group::Membership& membership)".
Group needs to interpret the three different cases from the return value:
1. retryable error -> retry
2. nonode -> None
3. non-retryable error -> Failure

It turns out it's not that hard and we have precedence in the code base: 
https://github.com/apache/mesos/blob/190e87c51d25646fa501ffca0bf7150157982050/src/state/zookeeper.cpp#L427

So I revised the review to have Group return a None in the case of NONODE.


- Jiang Yan


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/25663/#review53558
-----------------------------------------------------------


On Sept. 17, 2014, 11:12 a.m., Jiang Yan Xu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/25663/
> -----------------------------------------------------------
> 
> (Updated Sept. 17, 2014, 11:12 a.m.)
> 
> 
> Review request for mesos and Ben Mahler.
> 
> 
> Bugs: MESOS-1392
>     https://issues.apache.org/jira/browse/MESOS-1392
> 
> 
> Repository: mesos-git
> 
> 
> Description
> -------
> 
> See summary.
> 
> 
> Diffs
> -----
> 
>   src/log/network.hpp fc85a57a38f89190fe246f16cd1fde3168d70613 
>   src/master/detector.cpp 6436b8ee7e1ab6451a6b999a1cfbb2f79190e6ca 
>   src/tests/group_tests.cpp 7ed98956181f167e16cd723c049738f1c217c73b 
>   src/zookeeper/group.hpp 16f9b7b390551402e3c1eddaf5657aa18766b47c 
>   src/zookeeper/group.cpp 58491c01052b68ddaee6af32f33192d5a1f20e58 
> 
> Diff: https://reviews.apache.org/r/25663/diff/
> 
> 
> Testing
> -------
> 
> make check.
> 
> 
> Thanks,
> 
> Jiang Yan Xu
> 
>

Reply via email to