Debian Install and Hard Drive Debian Install

I followed the directions at

    http://elinux.org/BeagleBoardDebian

(except using ext2 for rootfs) and all went well until final boot when
I got

/dev/mmcblk0p2: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.
        (i.e., without -a or -p options)
fsck died with exit status 4
failed (code 4).
An automatic file system check (fsck) of the root filesystem failed. A manual f!
The root filesystem is currently mounted in read-only mode. A maintenance shell.
Give root password for maintenance
(or type Control-D to continue):

Pressing Ctl-D causes a reboot with the same result. I then changed
the bootargs to

   setenv bootargs 'console=ttyS2,115200n8 root=/dev/mmcblk0p2
rootwait rootfstype=ext2 rw'

and was able to boot. Question: Why does the original bootargs include
ro instead of rw?

Hey Ted

I followed the directions at

http://elinux.org/BeagleBoardDebian

(except using ext2 for rootfs) and all went well until final boot when

Don't use ext2, it will be a maintenance nightmare, specially with no
rtc battery... Do you have a rtc battery installed on your beagle?
It's not really worth the extra maintenance over ext3...

I got

/dev/mmcblk0p2: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.
(i.e., without -a or -p options)
fsck died with exit status 4
failed (code 4).
An automatic file system check (fsck) of the root filesystem failed. A manual f!
The root filesystem is currently mounted in read-only mode. A maintenance shell.
Give root password for maintenance
(or type Control-D to continue):

btw, is this lenny or squeeze?

Pressing Ctl-D causes a reboot with the same result. I then changed
the bootargs to

setenv bootargs 'console=ttyS2,115200n8 root=/dev/mmcblk0p2
rootwait rootfstype=ext2 rw'

and was able to boot. Question: Why does the original bootargs include
ro instead of rw?

ubuntu/debian need 'ro' vs 'rw' such that fsck will properly run at startup...

=====

Also, to increase longevity and possibly speed, I want to install onto
a bus powered usb hard drive. I understand that u-boot needs to boot
from the SD card, so the strategy is to put the kernel on the SD and
the rootfs and swap on the HD.

I tried the same installation as above, with the following changes:

1. During netinstall, format and install to the HD instead of SD. The
rootfs will be /dev/sda1
2. Change the ultimate bootargs to

setenv bootargs 'console=ttyS2,115200n8 root=/dev/sda1 rootwait
rootfstype=ext2 rw'

That's fine: I boot mine with: "console=ttyS2,115200n8 root=/dev/sda1
rootwait rootfstype=ext3 ro"

During (1) I had some hiccups, I never got to tasksel, although the
base system seemed to have installed. After (2), booting got pretty
far, but had several errors

Activating swap...[ 32.337860] uncorrectable error : <3>end_request:
I/O error, dev mtdblock0, sector 0
[ 32.345703] Buffer I/O error on device mtdblock0, logical block 0
[ 32.352172] uncorrectable error : <3>end_request: I/O error, dev
mtdblock0, sector 8
[ 32.360015] Buffer I/O error on device mtdblock0, logical block 1

Safe to Ignore, mtdblock0 is the nand... the net install for lenny is
an old 2.6.29 kernel, later versions get rid of that message.

pastebin your dmesg so we can see the rest...

In the end, booting got as far as starting cron, then hung.

Question: Has somebody had this setup before, and how should I deviate
from the stock debian elinux directions?

Yes, multiple installs over the last year. 3-4 beagles running 24/7
on Debian Lenny (dist-upgraded to squeeze)

Regards,

Robert,

Thanks for the help. I've got a zoo of CPUs for my crypto research,
and am very excited to get and ARMv7 machine, finally.

I reinstalled by following the instructions at http://elinux.org/BeagleBoardDebian
to the letter, except that I grabbed the Squeeze installer instead of
the Lenny one, and I formatted and installed to a USB HD at /dev/sda
instead of to the SD card.

Don't use ext2, it will be a maintenance nightmare, specially with no
rtc battery... Do you have a rtc battery installed on your beagle?

No rtc battery. Now ext3.

ubuntu/debian need 'ro' vs 'rw' such that fsck will properly run at startup...

Installation to /dev/sda1 went without a hitch. I setup my normal.cmd
as follows, followed by mkimage.

setenv bootcmd 'mmc init; fatload mmc 0:1 0x80300000 uImage; bootm
0x80300000'
setenv bootargs 'console=ttyS2,115200n8 root=/dev/sda1 rootwait
rootfstype=ext3 ro'
boot

pastebin your dmesg so we can see the rest...

When I reboot, it gets pretty far, but ultimately complains about fsck
failing and the rootfs being read-only. If I push Cntl-D, it cycles
and gives the exact same error. I then entered the maintenance shell
and manually ran fsck, after which a reboot was successful.

Thanks agian for your help!

-Ted

I’ve done the same but used ext3 - I have no issues so far, 14 days uptime…My Beagleboard is 15 days old :slight_smile:

I would think your issue is likely related to the use of the ext2 file system.