Re: Review Request 15099: Cleaned up the Registrar.

[Ben Mahler <be...@gmail.com>]

> On Nov. 7, 2013, 5:47 p.m., Vinod Kone wrote:
> > src/master/registrar.cpp, line 307
> > <https://reviews.apache.org/r/15099/diff/2/?file=377780#file377780line307>
> >
> >     How is this possible? i.e., update() is called when there are no mutations.

It's not possible, but I believe benh was looking to just do the right thing if this was called with no mutations, in which case update() is a no-op. I've changed the logic a bit here to return immediately in this case.

We could use a CHECK instead for now, I think either are fine.


> On Nov. 7, 2013, 5:47 p.m., Vinod Kone wrote:
> > src/master/registrar.cpp, line 193
> > <https://reviews.apache.org/r/15099/diff/2/?file=377780#file377780line193>
> >
> >     If there is a network partition with the zk are we going to wait forever? Are we assuming this is ok because the leading master would eventually commit suicide if there is a partition?

The storage operation will be stuck, the leading master will self-terminate at some point.

This TODO is also for pending operations for other reasons, we likely don't want to wait forever during normal operation if something gets stuck. However, using a fixed timeout is dangerous unless it determined to be much larger than any reasonable storage latency.


- Ben


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/15099/#review28384
-----------------------------------------------------------


On Nov. 5, 2013, 2:34 a.m., Ben Mahler wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/15099/
> -----------------------------------------------------------
> 
> (Updated Nov. 5, 2013, 2:34 a.m.)
> 
> 
> Review request for mesos, Benjamin Hindman and Vinod Kone.
> 
> 
> Repository: mesos-git
> 
> 
> Description
> -------
> 
> This is addressing my comments on https://reviews.apache.org/r/14383/
> 
> 
> Diffs
> -----
> 
>   src/master/registrar.hpp 5742c365bac0082efea53c35333d3c3503494066 
>   src/master/registrar.cpp 42fe30ededd004e380abc835c224649e02999f11 
>   src/master/registry.proto bd850997c57153cdfc0c39d0dae7d2b802034aa3 
>   src/tests/registrar_tests.cpp 51975f5aa2725f64d2a83387a381e5078c8b9e1d 
> 
> Diff: https://reviews.apache.org/r/15099/diff/
> 
> 
> Testing
> -------
> 
> make check
> 
> 
> Thanks,
> 
> Ben Mahler
> 
>