[Steve Langasek] > Setting up insserv (1.12.0-10) ... > info: Checking if it is safe to convert to dependency based boot. > error: Obsolete conffile /etc/init.d/gdomap left behind by package > gnustep-base-common > error: Problems running insserv: > insserv: script lvm: service lvm already provided! > insserv: There is a loop between service lvm2 and console-setup if started > insserv: loop involving service console-setup at depth 10 > insserv: loop involving service mountnfs-bootclean at depth 9 > insserv: There is a loop between service lvm2 and console-setup if started > insserv: loop involving service checkfs at depth 3 > insserv: loop involving service lvm2 at depth 2 > insserv: loop involving service udev at depth 1 > insserv: There is a loop between service checkfs and lvm2 if started > insserv: loop involving service mountall at depth 15 > insserv: There is a loop between service mountall and lvm2 if started > insserv: loop involving service networking at depth 25 > insserv: loop involving service console-screen at depth 30 > warning: It is not safe to convert to dependency based boot.
I suspect this loop is because of init.d scripts without or with incorrect LSB header, left behind from removed but not yet purged packages. Do you have both lvm and lvm2 init.d scripts? > It then pops up a debconf prompt, asking me whether I want to enable > dependency-based booting. > > Of course, since this is using the dialog frontend, it pops this up /over/ > the warning message saying that it's not safe to enable dependency-based > booting! > > The default answer for the question is set to 'no', but the question ought > to be suppressed entirely if there's only one right answer... especially if > the wrong answer is dangerous, as seems to be the case here. Note that even if the question is presented, if it isn't safe to convert, the 'yes' reply would be rejected (the postinst whould try to convert, conclude that it would be unsafe to do so and report that it is unable to convert). So the system would not be converted as long as a loop or other problem is detected. I agree that it do not really make sense to ask a question in this case when the answer will be ignored. Happy hacking, -- Petter Reinholdtsen -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
