From mboxrd@z Thu Jan 1 00:00:00 1970 From: ludo@gnu.org (Ludovic =?UTF-8?Q?Court=C3=A8s?=) Subject: bug#24943: Error when starting a vm-image in qemu Date: Sat, 19 Nov 2016 18:48:44 +0100 Message-ID: <87oa1b2web.fsf@gnu.org> References: <20161115050131.GA14613@khaalida> <871sydgimv.fsf@gnu.org> <20161118220908.GA10079@khaalida> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Return-path: Received: from eggs.gnu.org ([2001:4830:134:3::10]:55635) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1c89lD-0001ne-GG for bug-guix@gnu.org; Sat, 19 Nov 2016 12:49:08 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1c89l8-0005my-Ln for bug-guix@gnu.org; Sat, 19 Nov 2016 12:49:07 -0500 Received: from debbugs.gnu.org ([208.118.235.43]:48136) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1c89l8-0005mt-Ib for bug-guix@gnu.org; Sat, 19 Nov 2016 12:49:02 -0500 Received: from Debian-debbugs by debbugs.gnu.org with local (Exim 4.84_2) (envelope-from ) id 1c89l8-0000KH-CS for bug-guix@gnu.org; Sat, 19 Nov 2016 12:49:02 -0500 Sender: "Debbugs-submit" Resent-To: bug-guix@gnu.org Resent-Message-ID: In-Reply-To: <20161118220908.GA10079@khaalida> (dian cecht's message of "Fri, 18 Nov 2016 14:09:08 -0800") List-Id: Bug reports for GNU Guix List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: bug-guix-bounces+gcggb-bug-guix=m.gmane.org@gnu.org Sender: "bug-Guix" To: dian_cecht@zoho.com Cc: 24943-done@debbugs.gnu.org dian_cecht@zoho.com skribis: > I thought I had included that info, but yes that does work. Part of the r= eason > for filing the bugreport was so that this case could be handled more eleg= antly > instead of bugging out and dropping a user in the REPL. Unless one is a d= ev, the > REPL might as well be a kernel panic, only without the (potentially) usef= ul > timeout. > > I'd personally prefer it if the image generated by vm-image could either > > A) Spit out an error that there isn't a networking device found and shutd= own > (not a fan, but if there is a reason vm-image /has/ to have a working net= work > connection, this would likely be better. Including a chance to hit a key = to drop > a dev/advanced user into the REPL would be nice as well) or > B) Spit out an error and finish the booting process, with guix complainin= g about > no network connection any time it's needing to download something (easily= the > better option). Yes, that makes sense. On closer inspection, that feature (setting up QEMU guest networking directly from the initrd) is something we need in only 1 situation, which is =E2=80=98expression->derivation-in-linux-vm=E2=80=99, so commit 6129dd8b5989f77b2976c68ecdf1f7dbfa63ec46 removed that feature for images produced by =E2=80=98guix system vm=E2=80=99 and =E2=80=98guix system vm-im= age=E2=80=99. So now you should no longer have this problem when networking is missing. Thank you, Ludo=E2=80=99.