Bug#582177: initramfs-tools: fails to load ramdisk on boot; ... BOOT DEBUG DATA

[maximilian attems]

On Wed, May 19, 2010 at 02:50:42PM -0400, cropper at acm.org wrote:
> 
> 
> Ok, more data... old school capture (pen and paper!)

thanks a lot that made the picture much clearer.
 
> _____________________________________________
> Preparations:
> 
> 1) Reinstalled most recent initramfs-tools version
> 2) executed "update-initramfs -c -k all"
> 3) executed "update-grub"
> 
> _____________________________________________
> GRUB menu at boot:
> 
> linux /vmlinuz-2.6.32-3-amd64 root=/dev/mapper/md1_crypt ro debug=vc rootdelay=12
> echo Loading initial ramdisk
> initrd /initrd.img-2.6.32-3-amd64
> 
> _____________________________________________
> Kernel output (LITERAL DATA, except for "... area below"):
> 
> [1.185516] rtc_cmos 00:0a: setting system clock to 2010-05-19 18:18:27 UTC (1274293107)
> [1.185628] Waiting 12sec before mounting root device
> [1.199432] input: AT translated Set 2 keyboard as /devices/platform/i8042/serio0/input/input1
> [12.992053] List of all partitions:
> [12.992128] No filesystem could mount root, tried:
> [12.992221] kernel panic - no syncing: VFS: unable to mount root fs on unknown wn-block(0,0)
> [12.992287] Pid: 1, comm: swapper Not tainted 2.6.32-3-amd64 #1
> [12.992340] Call Trace:
> [12.992397] [<ffffffff812ed349>] ? panic+0x86/0x141
>     ...             ...                ...
>     ...             ...                ...
> [12.992880] [<ffffffff81011ba0>] ? child_rip+0x0/0x20

aboves says that linux-2.6 is not getting the initramfs.
thus indeed blaming GRUB seems the current state of affair.
 
> _______________________________________________
> Observations and opinions:
> 
> This output format really has NOT changed in the days we have been messing with this problem.
> 
> 1) the RAID is NOT starting
> 2) it never gets to the crypto (not started either)
> 3) no USB/firewire devices are detected (no evidence of drivers loading)
> 4) no block devices (Hard disks, CDROMs, etc) are detected (no evidence of drivers loading)

indeed 1-4 can't happen if initramfs is not unpacked.

strange that size seems not the issue:
ls -lrt img-2.6.32-3-amd64.*
-rw-r--r-- 1 maks maks 9458502 May 19 04:32 img-2.6.32-3-amd64.working
-rw-r--r-- 1 maks maks 9665786 May 19 04:32 img-2.6.32-3-amd64.broken

also qemu just unpacks fine the "broken" one and it's good for file too:
 file img-2.6.32-3-amd64.broken 
 img-2.6.32-3-amd64.broken: gzip compressed data, from Unix, last modified: Wed May 19 01:49:38 2010
 
> _______________________________________________
> 
> Questions:
> 1) Could this be grub?
> 2) Could "resynchronizing" md0 fix this problem?  I seem to remember doing something like this and the problem went away in March... :-/
> 3) Any other suggestions?

adding GRUB maintainers to Cc as this new bug report seems to be a
recurring trend. #578473 seems to be duplicate of that one.

GRUB maintainers can you advise?