On Sat, 26 Sep 2015, Jeroen Dekkers wrote: > > How does failing the upgrade solve anything? The upgrade should only > > fail if the failure of the service to start was because something in the > > upgrade itself was broken; this is rarely the case. > > I think it solves the problem of notifying the user that something > went wrong quite clearly. Not in the correct way, I agree with that,
It also makes recovering from the errror absolutely unfriendly. Imagine an update in a dist-upgrade across releases fail (I had that recently with udev because one system was still running the squeeze kernel and I failed to check before; this one was rather easy to fix actually, compared to others I’ve seen over time). dpkg is the first step, then apt, but apt-get -f install often wants to remove stuff you’d rather keep, and you can’t continue the earlier running dist-upgrade, and you must not reboot before this is resolved either, etc. No, package installations must not fail simply because a service failed to start or restart, with possibly (VERY) few exceptions that ought to be discussed on d-devel before being implemented (the udev one is actually a good one as it prevents major damage by making the system unbootable). On that note, “apt redo” would be cool to have for such cases… a mode to use when the last action failed, to retry it, without even needing to re-resolve dependencies and all… (yeah not going to be able to write the code, so not privileged to wish for it). bye, //mirabilos -- tarent solutions GmbH Rochusstraße 2-4, D-53123 Bonn • http://www.tarent.de/ Tel: +49 228 54881-393 • Fax: +49 228 54881-235 HRB 5168 (AG Bonn) • USt-ID (VAT): DE122264941 Geschäftsführer: Dr. Stefan Barth, Kai Ebenrett, Boris Esser, Alexander Steeg