Talk:Real NAS

You could simply use a IDE-CF-Card converter and add a 256 MB Flashdisk instead of the Notebook disk. Okay, 256 MB is not to much, but you can go bigger if you want.

The main subject on thinking about this is the heat developement by another disk...and maybe speed when booting...


 * Yes, a 256MB CF card should quite do the job if you are going to keep the original OS. But if you plan to install another distribution and don't want to flash a custom kernel you'll need to make a really small (about 4mb) hda1 partition, install just busybox the loader module and your kernel and boot form it to hda2 with your real root.
 * --Yvasilev 21:28, 12 Aug 2005 (CEST)

My Terastation is crashed because I copy install some component (from Debian web site) on it after the machine was hacked. I try to restore the disk firmware by mounting the disk and restore all image.dat disk firmware. Following is the operation and error message:

JONES:/t# mount -t xfs /dev/hdb1 /x mount: wrong fs type, bad option, bad superblock on /dev/hdb1, missing codepage or other error (could this be the IDE device where you in fact use      ide-scsi so that sr0 or sda or so is needed?) In some cases useful info is found in syslog - try dmesg | tail or so

where teh device 'hdb' are created from one of the original RAID disks (by 'dd' comand which is also introduced in this article).

Is there any other method to restore the firmware back to original fab setup? or the above method has something mistake that I didn't catch?

Thank you very much!


 * I don't quite understand what is what you mean by "where teh device 'hdb' are created from one of the original RAID disks (by 'dd' comand which is also introduced in this article)". The dd command can't create a device, it can only access (read or write) a contents of a device or a file. But if you created an image with

dd if=/dev/250gb_disk of=original-sectors-0-65.raw bs=8225280 count=66
 * you can restore it later with:

dd if=original-sectors-0-65.raw of=/dev/250gb_disk bs=8225280 count=66
 * you'll need to do it to all 4 of our disks so RAID1 will not recover to the previous broken state. Actually you probably may want to do it to just one HD and then try booting the Terastation with the other 3 disconnected)


 * Also please let us know what Debian packages caused this, as the ones I have tried worked without any problems.
 * --Yvasilev 23:01, 27 Oct 2005 (CEST)

Thanks for your reply. I found a way to recover my Terastation by cleaning all Raid 1 partition and applying the original firmware package. However, this will rebuild your Raid 5 disk array and destroy all data on the disks.

''Also please let us know what Debian packages caused this, as the ones I have tried worked without any problems. ''

My system was crash after I install w-get, GCC and SAMBA 3.x package. All of them are got form Debain web with PPC version.

By the way, I found the four State/Full led doesn't light again after installing SAMBA 3.x (I only copy the deamon and its library and so that it didn't crash my system). Is there any method to identify the root cause? I also found that the SAMBA 3.x is not stable because sometimes it doesn't allow me to access Terastation from Windows. Is there any a normal method to install SAMBA 3.x? Thanks in advance.

--jones 20:34, 28 Oct 2005 (CEST)

Setting up OS disk inside a running terastation
A better way to set up the OS disk ...

Shut down the box, add the new drive as master on DISK1 connector and change existing disk to slave Boot the box Log in as root Copy the partition table from another disk

dd if=/dev/hdb of=/dev/hda bs=512 count=1

I foud it necessary to delete partitions 3 and 4 as they seem to cause all sorts of SectorIDNotFound errors, I re-created sensible ones at the same time

mfdisk -c /dev/hda d   3 d   4 n   e    3 66   4864    w

It does, however, seem to result in the red fail light flashing for disk 1.

Add the new disk into the raid 0 array

mdadm --manage /dev/md0 --add /dev/hda1

Remove a disk from the array to force a rebuild

mdadm --manage /dev/md0 --fail /dev/hdb1

Watch it rebuild with ...

cat /proc/mdstat

Set up the swap partition

mkswap /dev/hda2

Reboot and check everything appears to be working, because if the new disk is a problem, it will be difficult to put the old array back together again as nobody seems to know the root password for EM mode.

You should now be able to remove all the other disks from the array with

mdadm --manage /dev/md0 --fail /dev/hd{c,e,g}1 mdadm --manage /dev/md0 --remove /dev/hd{c,e,g}1

And remove the unneeded partitions

mfdisk -c /dev/hdb d   1 d   2 dd if=/dev/hdb of=/dev/hdc bs=512 count=1 dd if=/dev/hdb of=/dev/hde bs=512 count=1 dd if=/dev/hdb of=/dev/hdg bs=512 count=1

And finally, reboot again.

This method will ensure that the data partitions are kept intact.

Some questions about this howto
Hi! Great howto! :-)

Before I start modifying our terastation to turn it into a Debian server, I would like to ask a few questions:


 * I plan to hack a 2 TB Terastation HomeServer. Is the bs= parameter for the dd command the same with this Terastation variant? Could this number be verified with the xfs_info command?
 * Why does the last dd command in section Using a Separate Computer use count=64 instead of count=66?
 * Any hints how to start a Debian installation after the hard disks are set up?

Thanks for your help! Maybe this discussion leads to some further enhancements of the howto. :-) --SteBo 16:47, 16 February 2007 (CET)