FS#27385 - [mkinitcpio] Cannot Determine Major/Minor for root partition

Attached to Project: Arch Linux
Opened by mark hadman (markhadman) - Thursday, 01 December 2011, 14:24 GMT
Last edited by Dave Reisner (falconindy) - Sunday, 04 December 2011, 03:19 GMT
Task Type Bug Report
Category Packages: Core
Status Closed
Assigned To Dave Reisner (falconindy)
Architecture i686
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 4
Private No

Details

Description:

After kernel upgrade to 3.1.3 can no longer boot the system, since it says cannot determine minor/major for the root partition. The output of "waiting for device <blah>" gets sucked into the $root device variable, so its looking for a root device of "waiting for device /devsda1 ... /dev/sda1" with a newline thrown in there for good measure.

Workaround:
The quick workaround is to add "quiet" to your kernel line in /etc/grub.conf. This prevents the "waiting for" message from being printed. So my grub.conf line that changed was

kernel /boot/vmlinuz-linux root=/dev/disk/by-label/archusb ro

to

kernel /boot/vmlinuz-linux root=/dev/disk/by-label/archusb ro quiet



Additional info:
* package version(s)
Linux 3.4.1-1
mkinitcpio 0.8.0-2

This task depends upon

Closed by  Dave Reisner (falconindy)
Sunday, 04 December 2011, 03:19 GMT
Reason for closing:  Fixed
Additional comments about closing:  mkinitcpio 0.8.0-3

backported fix from git
Comment by Dave Reisner (falconindy) - Thursday, 01 December 2011, 16:36 GMT
Uh, no. There must exist a LOGICAL explanation for this failure. How about booting with rootdelay=30 instead of your 'quiet' voodoo. If you're using a USB drive to boot, it's reasonable to believe that it might actually take longer than 10 seconds to show up.

You, of course, get a resounding 'cannot reproduce' from me across a half dozen machines.
Comment by mark hadman (markhadman) - Thursday, 01 December 2011, 17:13 GMT
Nice.

Adding rootdelay=30 to the kernel line fails to solve the problem. Adding quiet to the kernel line consistently solves the problem. I will here transcribe the error message complete with newlines:

ERROR: Unable to determine major/minor number of root device 'Waiting 30 seconds
for device /dev/sdb ...
/dev sdb1'.

Three other users with the same problem today:
https://bbs.archlinux.org/viewtopic.php?id=131163
https://bbs.archlinux.org/viewtopic.php?id=131226
Comment by Dave Reisner (falconindy) - Thursday, 01 December 2011, 17:45 GMT
Will be fixed in next release.
Comment by Andrew Gaydenko (student975) - Thursday, 01 December 2011, 17:54 GMT
I'm next here, details are described in the thread https://bbs.archlinux.org/viewtopic.php?id=131239
I confirm 'quiet' voodoo :-) does help and 'rootdelay=50' doesn't.
Comment by Dave Reisner (falconindy) - Thursday, 01 December 2011, 17:56 GMT
"Me too" posts are not helpful. Please do not leave them.

poll_device in init_functions needs to write to stderr, not stdout.

https://projects.archlinux.org/mkinitcpio.git/commit/?id=7effc762a2de
Comment by Andrew Gaydenko (student975) - Thursday, 01 December 2011, 17:56 GMT
"Architecture" can be set to both as far as I have x86_64 one.
Comment by Andrew Gaydenko (student975) - Thursday, 01 December 2011, 18:10 GMT
Dave, sorry for noise: by some magical occasion I have missed your post ("Will be fixed in next release"). It wasn't me who mentioned voodoo :) Probably I was during too long period in the textarea editing - there is nine minutes interval only between our messages.

Thanks!
Comment by Ted (skytod) - Friday, 02 December 2011, 12:09 GMT
Hey, if quiet doesn't work (print the same: Unable to determine major/minor number of root device), what should I do?
Comment by Dave Reisner (falconindy) - Friday, 02 December 2011, 12:18 GMT
Regenerate your initramfs or boot from the fallback. This bug does not apply to you (nor is what you have a bug).
Comment by Ted (skytod) - Friday, 02 December 2011, 12:36 GMT
Fallback is the same result and I can't regenerate initramfs without loaded system. I misunderstand about a bug that doesn't apply to me.
Comment by Ted (skytod) - Friday, 02 December 2011, 19:52 GMT
I'm not only one. One more "me too": https://bbs.archlinux.org/viewtopic.php?pid=1023892#p1023892
It's related bug.
Comment by Dave Reisner (falconindy) - Friday, 02 December 2011, 20:08 GMT
Spare me.

- Do not post "me too" as I noted earlier in this ticket, which you clearly did not read. Since you can't be bothered to read the FS, I can't be bothered to read your BBS link.
- Whatever borked image you've encountered is something that we see nearly every kernel release from a small subset of users. Whatever you might think, it is NOT related to this particular FS. Grab a live CD. Regenerate your initramfs.

This FS is solved as far as I'm concerned. I've only left it open for other people to find in case they have the same problem which has the solution of using 'quiet' or applying the patch I've posted.
Comment by Ted (skytod) - Friday, 02 December 2011, 20:50 GMT
Ok. I fixed it by chrooting and downgrading. I'm deeply sorry, but I was very angry when couldn't use the computer for a long time.

Loading...