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

format them in a computer or set the disk to format as a raid or some other then your are trying in now.

Success!

I started the format from another machine and it completed.  The machine that didn't work was running Vista Home Premium with IE7 & One Care, the machine that did work was running XP SP2 with IE7 and Norton.  Both had the popup blocker disabled.  The only notable difference was that on the XP machine I got an OK button after the format and on the Vista machine it stopped responding at 94%. 

I have just upgraded to 1.03 firmware and repeated the process and the same thing happens - works fine on XP/IE7 fails on Vista/IE7.  I will have a dig around and ee if I can find out why.

Cheers

Mik

I just responded to this same issue in a different post, so I'll summarize how I reformatted my pair of drives which had been configured as RAID 1 to a pair of individual disks despite this 94% error (and despite not having any other SATA hardware with which I could fdisk the drives)...

1. backup your config, then do a factory defaults restore.
2. remove both drives, and place the "left" drive in the right slot (leave the other disk out for now)
3. power-on and format as individual disk, then power-off
4. insert the other drive in the open bay, then power-on
5. format the "new" disk as an individual disk
6. et' viola!  both formats completed properly and the system was back in working order.

this is sorta off topic but helpfull with this kinda problem

I run SpinRite 6 on level 4 on all my drives before I use them to determine if there are any problems with the read/write

if there is a problem like a suspect sector it will prevent formatting by most means if the drive cant write to that sector of the drive and hang at that point (obviously thats not the problem here since others have had it stop at the same point, but...) and spinrite can swap these sectors out for spare ones and the drive will work fine

yea it takes like 20 hours to run spinrite on a 1TB drive but it is worth it to determine that the drive can read/write to every single sector on that drive, there is really no other way to find this out other than writing and having it not work

SpectralDesign:  your approach worked great for me.  I had initially configured the box for RAID1 but after reading through some of this forum I decided to go with 2 separate disks (+ rsnapshot).  But then I ran into the 94% issue.  So thanks for bailing me out.

I hit the very same issue head on and can shed some light on it now.

The old DNS-323 RAID1 750GB was getting full so I used this opportunity to upgrade to the brand new Seagate 702.11 1TB AS disks.

After pulling out the old disks, plugging in the new set, which was factory formatted, I have tried to build a RAID 1 and no surprise- the format stopped at 94%.
I know that some of you waited for extra 15 minutes but this is not necessary. Once the disks stop spinning you can pretty much pull the plug and declare the disk format unsuccessful.

I have tried everything that was recommended before- deleted the partitions, swapped the disks, restored factory defaults, used the reset button,  formatted the disks individually (again 94%) and the best I could get was two individual volumes after the 94% RAID1 format.

Being under time pressure I had no choice but to copy the funplug setup from the old RAID1 750 GByte array using a linux PC and one of the old 750GB mirrored disks. If you ever needed it, the following set of commands starts RAID from a single attached former enclosure disk and mounts it onto a Linux PC:

mdadm --create --verbose /dev/md0 --level 1 --raid-devices=2 missing /dev/sdb2
mount /dev/md0 /mnt/source

The D-link disk was attached to the PC as /dev/sdb, make sure that /dev/md0 is not in use on your Linux machine and /mnt/source has to be created first.

Then I simply samba mounted on the same PC the new D-link enclosure:
mnt //ip_address_of_the_DNS323/volume_1 /mnt/target

and copied the ffp directory from /mnt/source to /mnt/target plus the funplug script file.

After restarting the D-link enclosure and gaining access to the DNS-323 I have dismounted the second disk:

umount /mnt/HD_b2
umount /mnt/HD_b4

and recycled my PC command:

mdadm --create --verbose /dev/md0 --level 1 --raid-devices=2 missing /dev/sdb2

You cannot dismount /dev/sda since the funplug is running from it and it would need a modded script detailed elsewhere.

After restarting DNS-323 I had no more telnet access to the enclosure and after a bit of investigation I found that previously copied ffp directory and the funplug script have vanished. The biggest surprise however was awaiting on the status page were suddenly the two individual volumes disappeared and instead there was a degraded RAID 1 array.

So I copied the funplug setup again and restarted the appliance, got telnet access to this D-link box of tricks and found what I expected,
cat /proc/mdstat
revealed RAID1 running with only /dev/sdb2 in degraded mode.

Hence a simple fix:

mdadm --add /dev/md0 /dev/sda2

started the whole rebuild process:

Personalities : [linear] [raid0] [raid1]
md0 : active raid1 sda2[2] sdb2[1]
      974133312 blocks [2/1] [_U]
      [==>..................]  recovery = 13.7% (134029056/974133312) finish=141.2min speed=99157K/sec
unused devices: <none>


Note the rebuild speed of 100MBytes per second, you can get this if you pull out Seagate's default SATA1 jumper from the connector block at the SATA connected disk side.

I do not expect any more hickups at this stage although I lost access to the web interface. It does not bother me much, the interface will most certainly come back after the rebuild is finished and the unit was restarted.

In conclusion I have no idea why a couple of clean Seagate disks fail to create a RAID array but it seems to me that the only predictable path is to hack it from the command interface. And I strongly believe that this issue is a bug in firmware, which would be worth fixing.

-SD

Last edited by skydreamer (2008-07-08 15:54:05)

curmudgeon wrote:

Executive Summary: Resolved by changing firmware 1.5 to 1.4 and/or setting Config Inactivity Timer to 90 min
I changed both before the successful format, so I can't say which one (or both) is the shizbit.
I had changed firmware before without any success, so I'm leaning towards the Config Inactivity Timer...

- snip -

I can confirm this works for me. I tried 1.05 with 90min Config Inactivity Timer and it does not work. Reflash it back to 1.04 and with 90min Config Inactivity Timer, it works.

Running RAID 1 with 2x WD 1TB HDD now.

I had the 94% problem with IE (firmware 1.04), and the Config Inactivity Timer set to 90min.  Everything worked fine when I switched to Firefox.

I came across this problem, too, and found success (thanks to Correz' suggestion), and it did not involve rolling back or installing different firmware...

I posted my success story here: http://dns323.kood.org/forum/viewtopic. … 203#p30203