On 2013-02-15 19:04, Adam D. Barratt wrote: > On Thu, 2013-02-14 at 21:04 +0100, Andreas Beckmann wrote: >> After reading all the related bug reports again, it looks like lvm2 is >> actually intended to work well even without udev running and this has >> been implemented in wheezy via fixing the init script dependencies. So >> let us backport this change to squeeze. >> >> Unfortunately I don't know how to actually test that lvm2 works in a >> udev-free squeeze. But the installation problems are gone with this >> change. > > That's certainly a good start. At least some hint that it actually works > would be good though; maybe the maintainers (CCed) have some input here?
Maybe we can revisit this for the next oldstable update. There was no feedback from the maintainers, but Pierre Ynard (the reporter of #543163) confirmed (see #687761) that he once had a working squeeze setup with lvm2 and no udev by removing the initscript dependency. (He is tracking unstable nowadays and needs/needed some more changes for getting modules loaded, but that's the price for not using udev - and seems to be due to his special hardware setup.) This change to the initscript won't change anything for existing squeeze installations - either they already have lvm2 + udev installed or manually changed the conffiles to allow lvm2 without udev. But it would simplify my work running (sometimes weird :-) upgrade tests to find "interesting" corner cases - lvm2 has a large rdep tree requiring a lot of workarounds ... Andreas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
