https://bugzilla.suse.com/show_bug.cgi?id=1198066 https://bugzilla.suse.com/show_bug.cgi?id=1198066#c5 Knut Alejandro Anderssen Gonz�lez <kanderssen@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mfilka@suse.com Flags|needinfo?(kanderssen@suse.c |needinfo?(mfilka@suse.com) |om) | --- Comment #5 from Knut Alejandro Anderssen Gonz�lez <kanderssen@suse.com> --- (In reply to Thorsten Kukuk from comment #4)

Big question: is netconfig really required here?

By looking at the netconfig ntp code: it creates a new temporary NTP server list (does it have access to all input files if it is running inside the new system, while everything else is running in the inst-sys? Or are we have now an incomplete list?) and updates the currently running chronyd from the inst-sys with this list.

1. does this really do what is expected? 2. why do we need to "update" the running ntp server list of the currently running chronyd from the inst-sys at the end of the installation? 3. why do we run something, which calls many other tools and modifies running daemons in the inst-sys from a chroot system? Is this really safe?

Since this all works fine without netconfig (that's really a bad documented scripting mess) and we clearly want to move to use only native NetworkManager (why limiting ourself and customers by having netconfig in the middle, which means we have to re-implement and maintain all the supporting scripts a second time at our own and miss all new upstream features?).

My suggestion would be: check for the existence of netconfig, and don't call it if it does not exist. Else you have to touch the code again in a few weeks.

I completely agree, BTW mchf was checking the current state of ntp configuration since we moved to chronyd so I guess these checks could be part of his current work. -- You are receiving this mail because: You are on the CC list for the bug.