DSM-G600, DNS-3xx and NSA-220 Hack Forum

antovint wrote:

1) backup all data

However you go about it, do the above!

If it were me, and not having many packages as you describe, I would backup the data (*and verify that the backup is a good backup* - by restoring it somewhere and checking that everything is intact - many people over the years have probably made a backup to say one or more DVDs or CDs (but the media isn't that important) and then haven't been able to restore it at all or not restore 100% of their data), pull the drives connect them to a PC and delete all partitions - in effect restoring the drives to the state they were when you got them.  Then reinsert, format via web interface, then reinstall ffp and your packages etc and then restore your data.  The 323 does remember your drive serial numbers though, so maybe a factory restore with the pinhole button at the back or via web interface (assuming that either will remove the serial numbers from flash memory) wouldn't be a bad idea too.  I am sure there will be lots of other suggestions though, so wait a while before doing it - somebody may have a better suggestion.
HTH

I built a tool for doing this.  It is attached.  Place this script in the root of your Raid array (same place as fun_plug is), then run it.  You may want to look through the code, first.  Make sure you follow all the on screen instructions.

This is assuming you are using FW 1.04 or newer with the 1.04 or newer partition structure.  It is NOT for the 1.03 or older!!!

As always, back-up your data first, but this script is designed to give you 2 independant drives each with all your data.

---

Attachments:
break_raid_1.04_1.05.sh, Size: 2,336 bytes, Downloads: 247

No, it is based on the partition structure which was changed at 1.04.  I have not heard of any changes in 1.06.  The F/W level only applies for doing the updates to the config files, because their locations changed.  The actual process to stop the raid is the same.  Eventially I will make this into a package that does not require reboots, but that involves getting ffp running from either a USB source or from memory.  Then it would involve shutting down all disk dependant processes to allow the array to be unmounted.  I may use something similar to ffp reloaded, I do not know, yet.  Time is just not something I have right now. 

If you know linux a little, or just want to see how it works, open up the script and look at it.  I tried to comment it so you can see what is going on.

Yes, the left bay is from watcing the front.  The error you got is because you did not insert the drive back in before running the second part of the script.  So you actually did not zero out the superblock on the drive. 

Also, do not confuse Volume_1 with being a drive.  Volume_1 refers to a share.  That is usually the drive in the right bay (drive # 0), but it also could be the one in the left (in the case of swapped drives, or when booting off 1 drive only in the left bay).  Drive # 1 is in the left and Drive # 0 is in the right bay -- always.

HD_a4 and HD_b4 are the partitions used to store your config files, which may not be properly up to date if the procedure was done incorrectly.

Thanks bq and fordem.

I did this procedure:
- run the script
- shutdown using web admin tool,
- remove drive 1 (left bay),
- reboot,
- shutdown using web admin tool
- then re-insert drive
- run /mnt/HD_a2/break2

I'm sure about that.
Anyway, what do you think I've to do to repair errors, maybe it's enough running the second part of your script with both disk inserted?

Thanks in advance
Anto

I ran break2 with both disks inserted and I have the same message:

root@MyNAS:~# sh /mnt/HD_a2/break2.sh start
mdadm: Couldn't open /dev/sdb2 for write - not zeroing
root@MyNAS:~# kill process
rmmod: lltd: No such file or directory
$Shutting down SMB services:
$Shutting down NMB services:
Refresh Shared Name Table version v1.04
umount: /mnt/HD_c*: not found
Refresh Shared Name Table version v1.04

here the break2 I used:

#!/bin/sh
grep -w sdb /proc/partitions >/dev/null 2>/dev/null
if [ ! $? -eq 0 ]; then
echo "Insert drive and run program again..."
else
mdadm --zero-superblock /dev/sdb2
rm /mnt/HD_a2/break2.sh
do_reboot
fi

Tks bq!