> On April 24, 2017, 7:25 a.m., Qian Zhang wrote:
> > After this change is introduced, I think we may need to update some docs, e.g.,
in https://github.com/apache/mesos/blob/master/docs/persistent-volume.md, it is mentioned
that a persistent volume might be created by one framework and then offered to a different
framework subscribed to the `same role`, but now I think it is not just `same role`, instead
it should be the role itself and any its descendant roles.
>
> Jay Guo wrote:
> Good catch! I actually wonder if this is correct at all... need some further discussion.
cc @mcypark @bmahler
I've updated the description of [MESOS-7033](https://issues.apache.org/jira/browse/MESOS-7033)
to be sure we capture this. It does seem unfortunate to me that there is no explicit way to
bind a reservation/volume to a framework, but this is an existing problem so we should tackle
it separately. Filed [MESOS-7651](https://issues.apache.org/jira/browse/MESOS-7651).
- Benjamin
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/58509/#review172760
-----------------------------------------------------------
On April 18, 2017, 3:44 p.m., Jay Guo wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/58509/
> -----------------------------------------------------------
>
> (Updated April 18, 2017, 3:44 p.m.)
>
>
> Review request for mesos and Michael Park.
>
>
> Bugs: MESOS-7149
> https://issues.apache.org/jira/browse/MESOS-7149
>
>
> Repository: mesos
>
>
> Description
> -------
>
> Other than looking at reservation made for a role, now allocator also
> aggregates reservations of all its ancestor in the hierarchy.
>
>
> Diffs
> -----
>
> src/master/allocator/mesos/hierarchical.cpp 051f749dd5921a322ca930a042c31814616d38f9
>
>
> Diff: https://reviews.apache.org/r/58509/diff/1/
>
>
> Testing
> -------
>
> see next patch in chain
>
>
> Thanks,
>
> Jay Guo
>
>
|