Description of problem: Based on Bug 1554472#c37 the distro_importer will enable all RHEL8 repos. Yet certain users want to minimize the amount of automatically enabled repositories during testing to avoid tainting the system with packages from said repositories. Instead of starting from a system with all repositories enabled, they need a minimal base system. Since Bug 1554472 enables all repositories, Beaker should provide a way for users to disable repositories with a ks_meta variable.
Idea: Based on Jan's idea from Bug 1554472 comment 37 we could use the repoid in form of a: no_<repoid>_repos. In reality that would become: * no_Buildroot_repos * no_AppStream_repos etc. The only problem with repoid is, that it can be camelcase and contain hyphens. For example: * AppStream * ComputeNode-optional-debuginfo Should these ids be normalised? Kept as is?
With RCM's efforts to move to unified composes, I think this item is of less priority after beta. The core tho still stands: If we have BaseOS + AppStream imported as a single compose, Buildroot will still be shipped as a separate and enabling all of them seem to be not the best choice. I'll raise this topic on the mailing list in order to get more ideas how we can implement something which works for everyone.
(In reply to Roman Joost from comment #1) > Idea: > > Based on Jan's idea from Bug 1554472 comment 37 we could use the repoid in > form of a: no_<repoid>_repos. In reality that would become: > > * no_Buildroot_repos > * no_AppStream_repos > > etc. In the future we have no idea if the repos will be called BUILDROOT, AppStream, etc. I think you want to do something like OS.action.noun.object ie) rhel8.disable.repo.BUILDROOT or rhel7.enable.repo.Optional If the names of the repos change in the future, and this is code properly, it should be easy to extend the options. That way you don't end up ks_meta data hell :) > > The only problem with repoid is, that it can be camelcase and contain > hyphens. For example: > > * AppStream > * ComputeNode-optional-debuginfo > > Should these ids be normalised? Kept as is? IMO. P.
*** This bug has been marked as a duplicate of bug 1664750 ***