https://bugzilla.suse.com/show_bug.cgi?id=1189772 https://bugzilla.suse.com/show_bug.cgi?id=1189772#c2 Martin Wilck <martin.wilck@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |patrick.schaaf@yalwa.com Flags| |needinfo?(patrick.schaaf@ya | |lwa.com) --- Comment #2 from Martin Wilck <martin.wilck@suse.com> --- So you use direct kernel boot on the host and have no vmlinuz/no initrd image on your VMs? This is a very special case. I don't think we should try to put more "intelligence" into the initrd rebuilding logic like "there's no vmlinuz, so let's skip". What we should discuss is e.g. an environment variable or a filesystem flag indicating that intitrd rebuild should be skipped on certain conditions. Users could then set this flag to set the policy. My personal opinion is that there are too many packages requesting initrd rebuilds after updates. For example, coreutils updates rarely actually necessitate an initrd rebuild. Others in SUSE engineering see it differently.

Since a month or two ago, I notice that tumbleweed started to nevertheless build initrds again

Could you put this more precisely perhaps? Maybe check /var/log/zypp/history and /var/log/zypper.log for updates of suse-module-tools, rpm-config-SUSE, dracut, libzypp/zypper, and kernel packages? -- You are receiving this mail because: You are on the CC list for the bug.