Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[daveb21]

Hey guys,

Hoping for some advice here.

Got a DNS-345 where two of the four drives are on their way to the grave. Setup is RAID5. I don't believe the drives are totally dead yet as one is still reporting OK in its SMART test and the other is detectable by the DNS-345 but fails its SMART test. I've powered off the DNS-345 and reseated the drives just to make sure no connectivity issues have developed over time.

Now here's the problem - the DNS-345 can see the four drives but won't recognise they are / were configured in a RAID5 set and hence doesn't report any Volumes. The Disk Management page just has the "hey you, wanna setup a RAID config?" button. I've got a feeling that maybe the 2 drives both reported failures at the same time and now the DNS-345 is being stubborn and has marked the drives/RAID as a permanent failure rather than a temporary failure that is now resolved.

Is there any way to force the DNS-345 to recognise that the drives are / were already setup? I liken the situation to that of a RAID controller losing its config and having to reload it.

I've opened a support case with DLink and they said that their ShareCenter support team only works 9-5 and I might get a call back within 24-48 hours.

Or is my only option to hook up the drives to a PC and use the recovery tools as mentioned here and in various other posts.

http://forums.dlink.com/index.php?topic=41800.0

Appreciate any thoughts and advice and thanks in advance.

Cheers,
Dave

[ivan]

My understanding of RAID 5 says that when you have two dead disks at the same time you have a dead RAID array and there is not much you can do to revive it.

The other thing is that RAID 5 is only a glorified RAID 0 with parity embedded - it is still just striping and not mirroring as is RAID 1.  With that in mind you will have the same problem trying to recover your data as you would have with a RAID 0 array.

Your best course of action is to replace the problem drives, format the RAID array (I would go for two RAID ! arrays rather than RAID 5) and restore from your backup.  You do have a backup of your data I assume.

If you don't have a backup then be prepared to spend a lot of money at a specialist data recovery company and even then they may not be able to recover it all.
 

[JavaLawyer]

Following up on Ivan's comments. . . As a last ditch effort, you can try using the recovery software mentioned in the sticky threads in this forum with the hope that the RAID array is intact, but falsely reporting a failure by the DNS-345 for some unexplicable reason.

[daveb21]

As I said I don't believe the 2 drives have suffered a permanent failure, just a temporary one however the Dlink device has "marked" the drives as unformatted and won't accept any evidence to the contrary.

Spoke to Dlink support and they suggested restoring the config file from backup. I did have a backup of it but I can't seem to find it.

The only parts of the config file missing are the RAID config and the Volume Config which I'm pretty sure I can recreate as it was setup quite simply. My understanding is the configuration file is simply a text or XML file so my plan is to export the configuration file as it stands now and manually enter the RAID and Volume config into the text file and hope that does the trick.

Has anyone got a config file they are willing to share with me - with all specific IPs / usernames / passwords blacked out. Or a primer as to how the config file is laid out.

FYI: I've been in IT for 10+ years predominantly in hardware scoping and Linux roles so I'm very comfortable and familiar with mucking about with RAID, text files, Linux etc.

I've already booted up a Linux box ready to look at R-Studio restore but I'd rather exhaust all native options before I remove the drives from the Dlink and start mucking about with them in the Linux shell.

[JavaLawyer]

Spoke to Dlink support and they suggested restoring the config file from backup. I did have a backup of it but I can't seem to find it.

The only parts of the config file missing are the RAID config and the Volume Config which I'm pretty sure I can recreate as it was setup quite simply. My understanding is the configuration file is simply a text or XML file so my plan is to export the configuration file as it stands now and manually enter the RAID and Volume config into the text file and hope that does the trick.

Has anyone got a config file they are willing to share with me - with all specific IPs / usernames / passwords blacked out. Or a primer as to how the config file is laid out.

Call for Help!

Is there a DNS-345 owner with a RAID-5 configuration who is willing to share a copy of their Configuration Settings file that's scrubbed of personal data (IP address, usernames, passwords, email address, etc.) with forum member daveb21? Or better yet, an extract of the config file just containing the RAID-5 configuration and volume configuration settings?

[daveb21]

Thanks for the shout out JavaLawyer!

[daveb21]

I booted up the device with all the HDDs removed (making sure I knew which one came out of which slot) and exported the configuration.

Looks like it is simply a gzip'ed text file which I unzipped using WinRAR and took a look at the text file. Albeit with what look to be a bunch of Linux format character returns which don't show up nicely in Wordpad (too lazy to copy the file to my Linux box). It I were coding it thats what I'd do for a backup export - simply gzip the text file before presenting to the user for saving.

It looks like the appropriate areas of the config file might be:

<dsk_mgr>
      <hdd_config>
         <raid_auto_sync>1</raid_auto_sync>
      </hdd_config>
      <dsk_diagnostic></dsk_diagnostic>
      <volume_mapping>
         <volume_1>Volume_1</volume_1>
         <volume_2>Volume_2</volume_2>
      </volume_mapping>
      <smart>
         <test>
            <hd_a>Pass [2014/04/27 08:23:48]</hd_a>
            <hd_b>Fail [2014/05/01 17:52:35]</hd_b>
            <hd_c>Pass [2014/05/01 17:53:21]</hd_c>
            <hd_d>Pass [2014/04/27 08:17:43]</hd_d>
         </test>
         <hd_a_sn>WD-WMC1H0096199</hd_a_sn>
         <hd_b_sn>W240B882</hd_b_sn>
         <hd_c_sn>W2F08WWM</hd_c_sn>
         <hd_d_sn>WD-WMC1H0065319</hd_d_sn>
      </smart>
   </dsk_mgr>

I'm gonna put the drives back in to the unit now and see if booting it up with no HDDs has caused it to rethink its opinion of the RAID status and try and detect the config off the drive signatures. Might be a long shot but I think it's worth it.

Next step after that is removing all 4 Production drives and replacing with 2 drives I've got lying around, configure them in a RAID1 set, export the config and hope I can reconstruct the RAID and Volumes portion of it from the config backup (substituting RAID5 and more disks as appropriate). I'll make sure I use Linux to edit the file

I'm also gonna have a look around and see if there is a guide to the options in the config file. Perhaps someone has cracked it open before, made changes and documented it. Or DLink provide a guide (unlikely).

Still..... sing out if anyone can help with a pre-existing config either full or just the relevant portions.

Cheers,

[daveb21]

Just an update in case anyone else finds themselves in the same position.

After a significant amount of trial and error with different methods and problems I'm currently in the process of recovering the data and here is the method I used (some of which is lifted from http://qnapsupport.net/?page_id=596):

1. Connected all drives to a Windows box
2. Used Seagate Seatools to scan both Seagate drives - one drive is completely dead and one drive had some bad sectors to which I applied the "fix bad sectors" option and it seems to have helped
3. Fired up R-Studio
4. Added the remaining 3 good drives to a Virtual RAID and set one drive as missing. It is important that the drives are setup in the correct order otherwise the RAID is out of order and any detection won't work. I used the config backup from the DNS-345 to remind me in what order the drives were setup. For me it was disk 2 that was dead so I set disk 2 as "missing"
5. Set the RAID Parameters as:

• RAID Type: RAID5
• Block order for: RAID5
• All other items left as default

6. Complete steps 10,11 and 12 from the QNAP Support link above. For 3x 2TB drives it took about a day. The key is during the scanning process if the scanning position block continues to increase it is working. If it stalls then one of the drives is really dead (I found this when I had the dead drive in the RAID set)
7. Select the "RecognizedXXXX" item in the Device View that reports as an ext4 filesystem and select "Open Drive Files"
8. Select the files to be restored and select a restore location as per steps 14-17 from the QNAP Support link above
9. Wait for files to restore (this is where I'm up to at the moment)

I still don't know why the DNS-345 wouldn't detect the RAID5 setup given 3 drives were still operational. But at least it looks like I'll get my data back.

This episode along with the 500 - Internal Server Error I've experienced when attempting to save the config file has shaken my faith in the product a bit. But I suppose overall it was very easy to setup and has just worked since day 1 and thats why I changed from using a Linux box with NFS / Samba to export my data....

Hope this post helps someone in a similar situation!

[JavaLawyer]

We also have R-Studio posted as a data recovery option here: DNS-345 - Data Recovery (Windows PCs), although the stepwise instructions you provided are very useful (and something I haven't yet included in one of our FAQs), thanks!

[ivan]

Hi Dave,

I must congratulate you on getting your data back without having to call in the specialists. 

There is one thing I would advise - DO NOT use the disk that has bad sectors in your array.  There are two reasons why not, first, the remapping of faulty sectors should be automatic (the firmware does this) and if you had to do it manually then there is a big problem, and second, another problem with this disk could bring your array down again.

Another thing, use the disk manufacturers disk tools to test the other drives as well, just in case (we have had to return too many drives to our suppliers in the past that now every drive we use is given a 7 day soak test and then tested with manufacturers disk tools - those that don't pass are returned for replacement).