Open Stack

On Mon, Nov 26, 2012 at 2:33 PM, Monty Taylor <mordred at inaugust.com> wrote:
> On 11/26/2012 12:00 PM, Matt Joyce wrote:
>>
>> I think RedHat is heavily involved in supporting OpenStack and it
>> behooves us to treat them well because of that.  I know they aren't
>> always the easiest distribution to work with but that's also their
>> strength.  But they are a bar worth raising ourselves up to.
>>
>> Of course, I am driven to say that if someone, such as say SuSE grew to
>> be more involved I would of course suggest we extend them the same
>> courtesy.
>>
>> The way I see it we'll never be able to avoid RHEL if openstack intends
>> to operate in enterprise.  It's going to happen whether we want it to or
>> not so we may as well embrace it.  But, we should let RedHat take care
>> of the bits that really are their own concern.  And I say the same of
>> SuSE.
>
>
> ++
>
>
>> Beyond that, everything that isn't ubuntu / debian is edge cases at the
>> moment.

+1.

>
> ++
>
> (I think that related to the above, it's on RedHat and Canonical both to
> take care of things like backporting latest library deps to their platform)
>
>
>> Windows and HyperV products for instance.  Or any potential port to a
>> BSD or Solaris.  Those are big questions.  Especially since the HyperV
>> team has contributed so much to OpenStack and has done incredible work.
>> Also because groups such as CERN who have been high profile and heavy
>> contributors/supporters are relying on that contribution.  But I feel
>> like at this point any work done there without significant user interest
>> or adoption should be done at the expense of interested parties.  And
>> that means keeping an open mind, and an open ear but not jumping through
>> hoops to keep them in the proverbial engineering loop.
>>
>> Opinions expressed are my own.
>>
>> And maybe that's part of the question.  We can discuss to what extent
>> this is an engineering concern.  But this also seems to fall at least
>> partially under the purview of the OpenStack foundation board.
>
>
> I think the main engineering question boils down to "when (do we|can we)
> stop caring about python 2.6" ... I don't think the day is today, because I
> do think it would be unduly negative towards RedHat at the moment. Other
> than that, I don't think us "supporting" one distro over another has a ton
> of quantifiable engineering effect.

There in lies the rub.  The length of time that RHEL 6.3 will be the
latest offering from RedHat could be longer than anyone would like to
support python 2.6.

I do think it's fair to put a line in the sand and say by 'h' or what
have you that python 2.7 will be the minimum. grizzly is certainly too
soon.

Regards,
Tom