5:06 p.m.
On Wed, Aug 24, 2016 at 08:57:52PM +0530, Kushal Das wrote:
* If we rename how to make sure that the users know that we are not
abandoning the cloud?
I've got a three-part answer to this part.
First, this absolutely *is* a refocus on container tech for scale-out
computing. That pretty much generally means "cloud taken for granted".
Second, Fedora Cloud Base never really caught on. I know that there are
some dedicated and serious users, but most people I'm aware of are
using it as an easy way to spin up a minimal Fedora VM (and it's the
_only_ way we provide an official vagrant image). That's very useful,
but it wasn't the goal. There *are* people I'm aware of using it for
actual cloud computing, but it's not taking the world by storm.
Third, we're already in a kind of concerning state, with F25 cloud
images not working since June and then here's the alpha deadline
looming. We need to figure out something else about that *anyway*.
* If we create a new group, then who else are still interested about
helping out the Cloud WG?
Generally, we have WGs for the editions and SIGs for spins and so on.
Those things aren't unimportant, just less formal. And we,
theoretically at least, hold the WGs to higher standards (like the
requirement to produce PRD updates).
* We will require new user stories related to Atomic. This also
brings
in the question about old user stories in current PRD. They are still
mostly valid. PRD updation is surely one major point towards Atomic WG.
Yes — nothing to add here.
* Who will maintain Atomic workstation? If it is workstation wg,
then
how to reduce duplication of efforts?
I'm unclear if Atomic Workstation will actually be actually based on
Project Atomic, be ostree and some similar technologies, but not really
connected. In the latter case, I think that would be "one more thing
named Atomic" in a confusing way, and it'd be better for that to have
another name.
But, if it is based on Atomic, I think we still have a clear
separation, because as we discussed at the FAD and after, the plan here
is to focus on what Josh and Adam dubbed "Project FAO" — a cluster
solution and possibly full-on OpenShift. In that case, Fedora Atomic Host
is a building block that could be shared and worked on by both the
Atomic WG (or FAO WG, or whatever) and the Workstation WG.
* What will happen to the Cloud Base image?
In my conception, there would still be a Cloud SIG, responsible for
cloud image technologies. They'd help both the Atomic WG and Server WG
produce and upload images to various cloud providers (where currently
"various" is Amazon EC2 and Digital Ocean (which is awesome) because
others are stalled on legal).
Given what I've said above about my impression of common use of the
current Fedora Cloud Base image, I think that artificat might well be
replaced with a Fedora Server Minimal Install Cloud Image (or whatever
terminology)
* Who will maintain the Vagrant images?
If the answer to the above plays out like I expect, it seems to be
well-aligned with Fedora Server. But Cloud WG could still help with
testing and etc.
* What will happen to our effort to push Fedora to different public
cloud services? We have AMI(s) right now, we also push to Digital
Ocean.
This effort is stalled and I don't have much hope. I think we might
have better luck coming at it from a different direction anyway.
I hope all of that makes sense!
--
Matthew Miller
<mattdm(a)fedoraproject.org>
Fedora Project Leader