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

Hello all,

I have discovering the DNS-323 and this forum for about a month. I was able to find a lot of useful information so far on the different subjects but right now, I am stuck and wife starts to be mad at me for spending so much time on it. I was hoping that some of you can have some answers on this issue

Here is my configuration and problem :
- DNS 323 with 2 SAMSUNG Spinpoint F2 ecogrenn 1TB - 32MB cache (ref : HD103SI)
- Build in RAID-1 array successfully.
- Firmware 1.07
- Install fun_plug 0.5 to play with mediatomb

I had not properly set the email settings at first so there was any issue, I did not notice. Anyway at one time, I shutdown the NAS and when switching it on the last day, I got this message :

Right Hard Drive Has Failed

Sincerely,
Your DNS-323

Even if both leds were blue. I think at first at a false alarm and therfore manually enabled rebuilding of the array without removing the assumed disk. From then (10 days), the array is rebuilding continuously (synchronisation time decreasing up to 0, then starting again at about 180 minutes). Here are some output from dmesg :

SCSI error : <1 0 0 0> return code = 0x8000002
sdb: Current: sense key=0x3
    ASC=0x0 ASCQ=0x0
Info fld=0x638dadd8
end_request: I/O error, dev sdb, sector 1670229376
raid1: sdb: unrecoverable I/O read error for block 1668140800
SCSI error : <1 0 0 0> return code = 0x8000002
sdb: Current: sense key=0x3
    ASC=0x0 ASCQ=0x0
Info fld=0x638dadd8
end_request: I/O error, dev sdb, sector 1670229378
SCSI error : <1 0 0 0> return code = 0x8000002
sdb: Current: sense key=0x3
    ASC=0x0 ASCQ=0x0
Info fld=0x638dadd8
...

As you see, it seems the disk has actually some problems. I therefore set the disk as faulty with mdadm, removed it and e2fsck it, accept all changes and restart the raid construction without any more success.

I then removed the "assumed" faulty disk and checked it on my PC with samsung estool. The tool told me that the disk had no failures. I therefore reformated it at low level and added it back on the dns-323. It then started to format, got stuck at 94% twice (back to low level format again).

I then moved to firmware 1.08 with the same 94% problem, then found in this forum that fun_plug may be the cause. I then renamed the fun_plug file and successfully went through the formatting (a great thanks to the community !!!).

I then enabled manually the rebuilding of the array and got stuck again  in the bloody reconstruction loop forever. Either the GUI actions or the CLI mdadm gives the same. I am pretty sure that my vendor will test the disk before accepting to RMA it to Samsung so I do not have any damn idea left.

I, of course, can format both disks (after saving all my data), but I would llike to understand what happened before restarting an heavy process.

Any one of you has this kind of issue ?

Thanks in advance,

Djoubil.

PS : attached id dmesg output

---

Attachments:
IO_Error.txt, Size: 15,629 bytes, Downloads: 419