Knowledgeable or interested in release engineering? We need you!

[Colin Walters <walters verbum org>]

As part of Project Atomic, we're pulling together two new delivery vehicles for software: Docker base images, and rpm-ostree.

For Docker, we need to be sure we're producing base images in a timely and reliable fashion as input packages change.  Right now, they're generated via Anaconda, and then we extract the data out to a tarball.

For more information on this in the Fedora context, see:

https://fedoraproject.org/wiki/Changes/Docker_Container_Image

On the host side, we have rpm-ostree which composes packages into trees.  This is a separate model from Docker.   See:

https://fedoraproject.org/wiki/Changes/Docker_Cloud_Image

The current Fedora release engineering is oriented around RPM packages plus install media like the boot.iso, DVD, and Live CDs.  The latter images are only generated infrequently - typically they're not even respun after a release.

Conceptually, we're introducing higher order delivery models that need to both *move fast* and be reliable.  When the next Heartbleed comes around, we need to know that we'll be able to respin both Docker base images and the OSTree tree(s) quickly and reliably.  And that our users will be able to rebase their applications and servers.

So, are you knowledgable and interested in release engineering? We could use you!  See the above links for Fedora Changes which link to their respective codebases, and feel free to jump in!