Comments on: Howto install a debian (from scratch) on a Buffalo Terastation Duo

By: s4mb4

s4mb4 — Mon, 14 Dec 2009 18:26:14 +0000

in step $> mkimage -A arm -O linux -T ramdisk -C gzip -a 0×0 -e 0×0 -n initrd -d initrd.gz initrd.buffalo

if i am reading the blog correctly, it really should say;

$> mkimage -A arm -O linux -T ramdisk -C gzip -a 0×0 -e 0×0 -n initrd -d initrd.gz new.initrd.buffalo

it should be called new.initrd.buffalo
and that file, new.initrd.buffalo should be scp’d over at the same time you scp deb_init.tar over to /boot on the NAS.

that would do 2 things, not cause the initrd.gz.1 issues i described above and would make sense of the step described here;

$> cd /debian
$> cp /boot/uImage.buffalo /boot/new.initrd.buffalo /debian
$> mv /debian/new.initrd.buffalo /debian/initrd.buffalo

Then the process would be correct (at least in syntax) AND my NAS would not be hosed …

By: s4mb4

s4mb4 — Mon, 14 Dec 2009 17:20:05 +0000

should i be able to pull the disks, perform a low level format on a seperate machine and plug them back in to get it started from scratch again?

i have the hardware to be able to accomplish that….

By: sebastien

sebastien — Mon, 14 Dec 2009 16:39:07 +0000

the chmod is not a problem. ou have to check that you can further access /dev/sda and /dev/sdb devices .. by doing a fdisk on it for exemple. And check that the partition are ok (“p” to print).

mdadm is the command, not mdadmin

good luck Jim.

By: s4mb4

s4mb4 — Mon, 14 Dec 2009 16:33:46 +0000

that doesn’t quite work for 2 reasons.
1. echo “- – -” > /sys/class/scsi_host/host0/scan

scan has been chmod’d to 200…

2. mdadmin command does not exist in EM mode…

By: sebastien

sebastien — Mon, 14 Dec 2009 13:53:50 +0000

so we said :

* You reboot the nas with no disks
* you end up in EM mode
* you run the acp_commander with -o option to telnet on it
* you plug the disks back
* you type : $> echo “- – -” > /sys/class/scsi_host/host0/scan
* same for host1 (for disk 2)
* then you issue : $> mdadmin –assemble /dev/sdaX /dev/sdbX (with 1 < = X <= 3)
* then you mount your new array
* copy your previous init.buffallo and image.buffalo on your /boot

then reboot

By: s4mb4

s4mb4 — Mon, 14 Dec 2009 13:47:53 +0000

how does one “ask the SATA controller to rescan the disk”?
i assume this is done from the linux command line since the web interface is not running while the NAS is in EM mode…

By: s4mb4

s4mb4 — Mon, 14 Dec 2009 13:37:00 +0000

but at what point do I scp the NEW initrd.buffalo file?
before or after the first reboot?

By: sebastien

sebastien — Mon, 14 Dec 2009 13:03:53 +0000

For the initrd.gz.1, this just that wget does not overwrite your existing file (and that’s a good thing!), just change you working directory and don’t get confused with what to do with which file

By: sebastien

sebastien — Mon, 14 Dec 2009 13:00:39 +0000

Hummm …

You NAS is NOT bricked … Don’t worry.

Unplug the disks… restart your nas. Wait 4 minutes to have it in emergency mode. Then run the acp_commander too gain access to it.

replug the disk and ask the SATA controlle r to rescan the disks.

Enjoy !

By: s4mb4

s4mb4 — Mon, 14 Dec 2009 12:53:25 +0000

well, thats it for me, i have officially bricked my NAS. I think i see my error.

after i edited my linuxrc file with the steps above, even though the intructions did not tell me to, i scp’d the modified initrd.buffalo onto my NAS and rebooted it.

i think after i ran rm /sbin/init, the ORIGINAL initrd.buffalo should have been there still….
then, after the NAS comes back up in EM mode, i should move the modified initrd.buffalo into the /debian directory.

hopefully i can RMA the device with my vendor and try again…