11:48 a.m.
On 10/20/21 5:29 AM, Robert P. J. Day wrote:
predictably, i suggested fedora would be a better choice, but
that's
not an option given corporate choice of ubuntu for dev boxes. so the
question is, will that still provide a reasonably complete set of
container tools relative to fedora using just standard packages?
There's likely a MASSIVE difference in the QA between Ubuntu, and
Fedora/CentOS/RHEL. As in, both upstream CI and downstream
package-level testing is vastly more faithful to the actual
user-experience in Fedora/CentOS/RHEL.
I cannot speak to the distro-level testing in Ubuntu. But I would
hazard to guess that even Debian's downstream testing is very likely
superior to Ubuntu's (even if if only because it necessarily happens
earlier).
What I can say for sure is, the upstream podman Ubuntu CI test
environment is highly customized with bleeding-edge, out-of-distro
dependent packages. This is required even to get podman to build
properly. Whereas on Fedora, all dependencies are eventually hitting
the normal distribution channels. i.e. What you use in Fedora is very
close to what we test upstream.
Our upstream Ubuntu CI is really more of a very basic "Can it work"
check, as oppose to "Can it work, and function for regular users". That
part is left to Debian/Ubuntu downstream testing, which I have no actual
knowledge about.
--
Chris Evich (he/him), RHCA III
Senior Quality Assurance Engineer
My personal robot overlord told me to include this signature line.