You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by Ben Mahler <be...@gmail.com> on 2013/10/04 20:14:51 UTC
Re: Review Request 13757: Fixed an issue where the Master unnecessarily
sends a "Framework failed over" message when the scheduler driver retries
an initial failover re-registration.
> On Sept. 16, 2013, 6:22 p.m., Vinod Kone wrote:
> > src/master/master.cpp, line 1899
> > <https://reviews.apache.org/r/13757/diff/1/?file=344097#file344097line1899>
> >
> > From MESOS-488 it sounded like we are in this situation when the reregister message is a duplicate. Are there other cases? If it is a duplicate can't we just ignore them? For example, can we ignore the message in reregisterFramework() if (framework->active && from == framework->pid)?
> >
> > In either case, please add a comment in the code.
Added a comment in reregisterFramework() as to why we cannot attempt to ignore duplicates.
Added a comment in failoverFramework as to why it is safe to not send FrameworkError when the pid is the same.
- Ben
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/13757/#review26147
-----------------------------------------------------------
On Aug. 23, 2013, 4:15 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/13757/
> -----------------------------------------------------------
>
> (Updated Aug. 23, 2013, 4:15 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Vinod Kone.
>
>
> Bugs: MESOS-488
> https://issues.apache.org/jira/browse/MESOS-488
>
>
> Repository: mesos-git
>
>
> Description
> -------
>
> See MESOS-659.
>
>
> Diffs
> -----
>
> src/master/master.cpp d53b8bb97da45834790cca6e04b70b969a8d3453
> src/tests/fault_tolerance_tests.cpp 10e52c401476eb8416361de49b8e4061bb7ac4f3
>
> Diff: https://reviews.apache.org/r/13757/diff/
>
>
> Testing
> -------
>
> Added a test that catches this case when not fixed.
>
>
> Thanks,
>
> Ben Mahler
>
>