Bug#356850: initramfs-tools: upon udev upgrade, system becomes unbootable if using lilo
when udev gets upgraded, the latest initram is regenerated via udev's
postinst. If using lilo, the system becomes then unbootable due to
lilo needing to get rerun in order to ack the changed initrd. See the
irc snippet included below for details.
04:15 <dato> maks: ok, that unbootability problem I had, had seemed to go away,
but came today, and this one time I noted the error
04:19 <dato> maks: initramfs-tools and linux-image updated to sid last week, no
updates since then; and, like the other time, seems like the udev update
04:20 <dato> maks: (initramfs-tools 0.53c, linux-image 2.6.15-8, udev went 0.086-1 -> 0.087-1 today)
04:21 <dato> maks: the 4 last lines when booting (and remember, no shell afterwards) were:
04:22 <dato> invalid compressed format (err=1)
04:22 <dato> VFS: Cannot open root device "301" on unknown-block(3,1)
04:22 <dato> Please append a correct "root=" option
04:22 <dato> Kernel panic - not syncing: VFS: unable to mount root fs on unknown-block(3,1)
04:22 <dato> maks: (using lilo, btw)
13:55 <maks> dato: sorry for not having your setup in my head, do you have a separate boot partition?
13:56 <maks> dato: is your root full?
14:16 <dato> maks: (np.) no, / is /dev/hda1, reiserfs, 275mb used, 425mb free.
usr, var, tmp and home are separated (in LVs, but that's irrelevant here)
14:23 <maks> dato: you are using lilo?
14:23 <dato> maks: yes
14:23 <maks> ah that explains it.
14:24 <maks> "invalid compressed format (err=1)" helped
14:24 <dato> yeah, that was probably there the previous time, but I didn't notice it.
14:24 <maks> udev upgrade runs update-initramfs which generates a newer initrd.img
14:24 <maks> and lilo doesn't grog that
14:26 <dato> oh hum. so somebody should really be running lilo, then? (this sounds like a serious bug to me)
14:26 <maks> care to bug report against initramfs-tools
14:26 <maks> need to think a bit about how to solve that
14:26 <maks> you can send your aboves irc log.
14:26 <dato> ok, will do.
14:27 <maks> dato: thanks, udev and evms are currently the only hook user of initramfs-tools
14:28 <maks> but the hooks of lvm2 and md will need to go to their packages too
14:28 <maks> cryptsetup hook for crypto root is in the works too
14:28 <maks> dato: so i need to find a good solution for that. :)
Adeodato Simó dato at net.com.org.es
Debian Developer adeodato at debian.org
An economist is an expert who will know tomorrow why the things he
predicted yesterday didn't happen today.
-- Laurence J. Peter