On Tue, Feb 26, 2019 at 2:35 PM <ivan.kabaivanov(a)gmail.com> wrote:
Hi,
I'm trying to get ovirt working on Linux From Scratch. So far so good, I can build
and install it, and engine-setup also works fine. Now, I'm at the last stage,
configuring it. And this is where I hit serious trouble: ovirt-host-deploy requires
dnf/yum/rpm.
It does not _require_ them. If they are missing, it won't do package
management. You can also force it to disable package management by
enabling the offline packager, check the README.
But since I'm not running a redhat-based distro, I have none of
those.
Short of reading the source code, what are the steps I need to manually perform to get a
working host?
There were efforts in 3.5 days to port oVirt to gentoo, and later some
efforts to port to Debian. You might want to check about these.
You can try checking the dependencies of ovirt-host and make them
available somehow (build from source or whatever).
You can also check the developer pages on the site, vdsm in particular.
And of course, you can simply try and see what happens. I do not think
we have documentation on using host-deploy from a dev-env engine
without installing from rpm, but you might find notes by previous
porters.
May I ask what you want to achieve? Personally I think that if I'd
wanted to port such a big project to an unsupported OS, I'd start by
installing on a supported OS, so that I can compare and check how
things should work where they do work. You can do this with
nested-kvm, no need for separate hardware.
Good luck and best regards,
--
Didi