FESCO will be discussing
https://fedoraproject.org/wiki/Features/FirstClassCloudImages and I figured
it would be helpful to have as much lined up before than as possible.
There's some discussion in a FESCO ticket, which this email expands upon.
(
https://fedorahosted.org/fesco/ticket/1008).
From a release engineering perspective, the key needs are:
* Automatic weekly scratch builds for rawhide and for the F19 branch once it
occurs. These should automatically use the latest kickstart from the
cloud-kickstarts git repository, and the resulting qcow2 tar.xz'd raw and
will go onto
alt.fedoraproject.org. Ideally, this will all be scripted
rather than manual. (Right now, one kickstart script per arch; in the
future, we may want to have test versions of multiple different
kickstarts, so the script should be a little flexible.)
At FUDCon, Dennis mentioned that he was planning to do this for the
install image, and this is only incrementally more than that, sot his
should be covered.
* Continue what we're doing with EC2. If we can get the Marketplace legal
issues settled, we'll only need to upload to US East and they'll do the
rest, which will decrease work, but otherwise, it's the same as we've been
doing, except a bit more formality with the alpha and beta builds.
* And then the qcow2 and raw.tar.xz builds for Alpha, Beta, and Final. These
should go onto the mirror network along with the install ISOs and etc., and
should have GPG-signed checksums. This probably needs at least some manual
intervention, but isn't completely unlike the existing EC2 image process.
Is there anything I've missed, here? Are there parts of this which will
require significant effort which I'm overlooking? What parts can *I* take on
directly?
As we're developing the updates to Koji to integrate a new image build
process, are there functional needs from the release engineering point of
view that need to be specified? I know Dennis *really* doesn't like the idea
of running ImageFactory and Oz at the builder level, but based on the
conversations with Ian McLeod and Jay Greguske, these tools are actually
_meant_ to work that way, and I think we can satisfy the functional needs --
but it'd be helpful to have the functional needs spelled out.
Thanks!
--
Matthew Miller ☁☁☁ Fedora Cloud Architect ☁☁☁ <mattdm(a)fedoraproject.org>