Open Stack

This seems a fair way of doing it.
Gives users and developers a fair warning.

If its not tested, because we will keep breaking it by accident, and
thats not fair on anyone, so those drivers shouldn't be in the main
tree

I wonder if we should requiring a certain level of testing or unit testing?
Like % coverage, or start VM, attach volume, delete VM?
Either way, this seems a good step forward.

John

On 10 July 2013 15:33, Russell Bryant <rbryant at redhat.com> wrote:
> On 07/10/2013 08:05 AM, Joe Gordon wrote:
>>
>>
>>
>> On Tue, Jul 2, 2013 at 7:38 PM, Russell Bryant <rbryant at redhat.com
>> <mailto:rbryant at redhat.com>> wrote:
>>
>>     Greetings,
>>
>>     Nova includes various compute drivers today, but the test coverage they
>>     receive varies quite a bit.  This is documented on the following wiki
>>     page.  The drivers are broken up into groups A, B, and C.
>>
>>         https://wiki.openstack.org/wiki/HypervisorSupportMatrix
>>
>>     We have two new compute drivers in the queue for Havana: docker [1] and
>>     z/vm [2].  I'd like to propose as a piece of criteria for inclusion that
>>     new drivers go into groups A or B.
>>
>>     Further, I would like to see *all* drivers move into groups A or B by
>>     the release of Icehouse.  I've been told that this is already in the
>>     works for VMware and baremetal, at least.
>>
>>
>> And if a driver doesn't go into group A or B by Icehouse?  Do we put
>> clear warnings in the release notes that they are 'use at your own
>> risk.' Perhaps if drivers stay in C by the end of Icehouse Development
>> we remove them, as we have no way of guaranteeing they work and shipping
>> broken code looks bad.
>
> Yes, removing them is what I had in mind.
>
> --
> Russell Bryant
>
> _______________________________________________
> OpenStack-dev mailing list
> OpenStack-dev at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev