Open Stack

Excerpts from Joshua Harlow's message of 2015-11-06 11:11:02 -0800:
> Clint Byrum wrote:
> > Excerpts from Doug Hellmann's message of 2015-11-06 10:28:41 -0800:
> >> Excerpts from Clint Byrum's message of 2015-11-06 10:12:21 -0800:
> >>> Excerpts from Dan Smith's message of 2015-11-06 09:37:44 -0800:
> >>>>> Worth mentioning that OpenStack releases that come out at the same time
> >>>>> as Ubuntu LTS releases (12.04 + Essex, 14.04 + Icehouse, 16.04 + Mitaka)
> >>>>> are supported for 5 years by Canonical so are already kind of an LTS.
> >>>>> Support in this context means patches, updates and commercial support
> >>>>> (for a fee).
> >>>>> For paying customers 3 years of patches, updates and commercial support
> >>>>> for April releases, (Kilo, O, Q etc..) is also available.
> >>>> Yeah. IMHO, this is what you pay your vendor for. I don't think upstream
> >>>> maintaining an older release for so long is a good use of people or CI
> >>>> resources, especially given how hard it can be for us to keep even
> >>>> recent stable releases working and maintained.
> >>>>
> >>> The argument in the original post, I think, is that we should not
> >>> stand in the way of the vendors continuing to collaborate on stable
> >>> maintenance in the upstream context after the EOL date. We already have
> >>> distro vendors doing work in the stable branches, but at EOL we push
> >>> them off to their respective distro-specific homes.
> >>>
> >>> As much as I'd like everyone to get on the CD train, I think it might
> >>> make sense to enable the vendors to not diverge, but instead let them
> >>> show up with people and commitment and say "Hey we're going to keep
> >>> Juno/Mitaka/etc alive!".
> >>>
> >>> So perhaps what would make sense is defining a process by which they can
> >>> make that happen.
> >> Do we need a new process? Aren't the existing stable maintenance
> >> and infrastructure teams clearly defined?
> >>
> >> We have this discussion whenever a release is about to go EOL, and
> >> the result is more or less the same each time. The burden of
> >> maintaining stable branches for longer than we do is currently
> >> greater than the resources being applied upstream to do that
> >> maintenance. Until that changes, I don't realistically see us being
> >> able to increase the community's commitment. That's not a lack of
> >> willingness, just an assessment of our current resources.
> >
> > I tend to agree with you. I only bring up a new process because I wonder
> > if the distro vendors would even be interested in collaborating on this,
> > or if this is just sort of "what they do" and we should accept that
> > they're going to do it outside upstream no matter how easy we make it.
> >
> > If we do believe that, and are OK with that, then we should not extend
> > EOL's, and we should make sure users understand that when they choose
> > the source of their OpenStack software.
> 
> Except for the fact that you are now forcing deployers that may or may 
> not be ok with paying for paid support to now pay for it... What is the 
> adoption rate/expected adoption rate of someone transitioning there 
> current cloud (which they did not pay support for) to a paid support model?
> 
> Does that require them to redeploy/convert there whole cloud using 
> vendors provided packages/deployment model... If so, jeez, that sounds 
> iffy...
> 
> And if a large majority of deployers aren't able to do that conversion 
> (or aren't willing to pay for support) and those same deployers are 
> willing to provide developers/others to ensure the old branches continue 
> to work and they know the issues of CI and they are willing to stay 
> on-top of that (a old-branch-dictator/leader may be needed to ensure 
> this?) then meh, I think we as a community should just let those 
> deployers have at it (ensuring they keep on working on the old branches 
> via what 'old-branch-dictator/leader/group' says is broken/needs fixing...)

Right, what I think where this leads though is that those who have
developers converge on CD, and those who have no developers have to pay
for support anyway. Running without developers and without a support
entity that can actually fix things is an interesting combination and
I'd be very curious to hear if there are any deployers having a positive
experience working that way.