• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Areca ARC-1210 - help! Incomplete RAID

C

CnnmnSchnpps

n00b
Joined
Mar 5, 2013
Messages
8
Looking to tap into your combined wisdom on this forum, please. Having some issues with my RAID5 volume 4x 2TB SATA on ARC-1210. Here's a rundown of the problem and steps taken to try and fix it so far:

- First symptoms were the raid volume being inaccessible under Windows. I could browse a few deep into the folder tree, but no files would open and media wouldn't play, etc.
- I tried to restart the PC but Windows hung forever on the reboot screen
- After a while I hard rebooted the PC
- Once it came back, I was getting a warning message on the Areca boot screen, and "Lost Rebuilding/Migration LBA" messages in log.

I had needed to do a level2reset previously, so I tried that first

In the repair section of the web interface do the following sequence
of commands, to the T and your array will show back up
#1) RESCUE
reboot like it tells you
#2) SIGNAT
#3) LeVeL2ReScUe (caps is important)
reboot like it tells you
#4) Array should be back now, do a SIGNAT again and itll stick
Click to expand...

This didn't work, so I tried the following instructions:

The event "Init LBA" happens when the background task has been interrupted (power down or reset) and that interruption can't be recovered. The unrecoverable problem could be a raid set disk member failure/missing while powers fail. This event is used for record the background task progress. In this case, if the interrupted condition can't be recovered any more. You can use the following procedure to clean the LBA record.1. Power off system then remove all disks while power is off.2. Power on system without any disk attached.3. After controller BIOS appears, press TAB/F6 to enter the McBIOS.4. Power off system, then reinserted all disks back while power is off.5. Power on system, the warning message should disappear.
Click to expand...

This didn't have any results either.

At this point, the raidset looks like below:

Raid Set IDE Channels Volume Set(Ch/Id/Lun) Volume State Capacity
Raid Set # 00
Ch01
Ch02
Missing
Ch03
IDE Channels
Channel Usage Capacity Model
Ch01 Raid Set # 00 2000.4GB ST32000542AS
Ch02 Raid Set # 00 2000.4GB ST32000542AS
Ch03 Raid Set # 00 2000.4GB ST32000542AS
Ch04 Free 2000.4GB ST32000542AS

Here are the log messages from around the time of the incident.. A bunch of "powered on" messages for some reason, then rebuilding, then lost LBA

2013-3-5 23:5:33 H/W Monitor Raid Powered On
2013-3-5 23:4:51 RS232 Terminal VT100 Log In
2013-3-5 23:4:33 000:000000163500 Lost Rebuilding/Migration LBA
2013-3-5 23:4:33 000:000000163200 Lost Rebuilding/Migration LBA
2013-3-5 23:4:33 Incomplete RAID Discovered
2013-3-5 23:4:32 H/W Monitor Raid Powered On
2013-3-5 23:3:50 RS232 Terminal VT100 Log In
2013-3-5 23:3:27 000:000000163500 Lost Rebuilding/Migration LBA
2013-3-5 23:3:27 000:000000163200 Lost Rebuilding/Migration LBA
2013-3-5 23:3:27 Incomplete RAID Discovered
2013-3-5 23:3:26 H/W Monitor Raid Powered On
2013-3-5 23:2:40 ARC-1210-VOL#00 Start Rebuilding
2013-3-5 23:2:40 Raid Set # 00 Rebuild RaidSet
2013-3-5 23:2:40 IDE Channel 3 Device Inserted
2013-3-5 20:5:11 H/W Monitor Raid Powered On
2013-3-5 20:4:29 H/W Monitor Raid Powered On
2013-3-5 20:3:47 H/W Monitor Raid Powered On
2013-3-5 20:3:5 H/W Monitor Raid Powered On
2013-3-5 20:2:23 H/W Monitor Raid Powered On
2013-3-5 20:1:41 H/W Monitor Raid Powered On
2013-3-5 20:0:59 H/W Monitor Raid Powered On
2013-3-5 20:0:17 H/W Monitor Raid Powered On
2013-3-5 19:59:35 H/W Monitor Raid Powered On
2013-3-5 19:58:53 H/W Monitor Raid Powered On
2013-3-5 19:58:11 H/W Monitor Raid Powered On
2013-2-21 0:13:34 H/W Monitor Raid Powered On
2013-2-21 0:12:51 H/W Monitor Raid Powered On
2013-2-21 0:12:9 H/W Monitor Raid Powered On

Would appreciate any guidance you guys can provide on how to possibly recover this volume! Please let me know if you need any additional info

Thanks!
-Dmitri
 
Those Seagates arent rated for RAID5 unless im mistaken? I did check on newegg they seem like normal consumer desktop drives.

Get a program for windows: HD Tune Pro (find the crack)
Connect the drive AHCI VIA SATA and read smart data, check for any warning if it is all fine then scan all sectors
 
If it is what I am expecting: this is what happens when a normal drive failure occurs and then tries to read from consumer grade drives which take longer than 7ms to respond. Hopefully that is the problem then we could just set them to 7ms VIA smartctl rebuild, and transfer data else where. But the timing is saying something else.
 
Hey guys thanks for weighing in! Not sure I fully understand the responses though, sorry relative n00b at this.

I thought they were RAID rated, but don't remember for sure. Will double check.

One thing I'm thinking is, since I have 3 drives under RaidSet #00 that still show up as members, and the other 4th one shows up as "free" - that 4th one could be the partially rebuilt drive. If I set that one to hot spare, it should recover automatically yes?

I'm not doing anything else until I hear back from Areca support and the folks on this forum, but that seems to be a logical approach yes?

May have to invest in some proper RAID-rated HDDs, maybe upgrade to 3TB at the same time.. Pricy tho

What's smartctl? the CLI interface?
 
I'm making some good progress towards recovering this volume. Over the weekend, I got 4x new 4TB drives and created a raw image of each disk. Analyzing the images in R-Studio using this procedure, I was able to determine following correct parameters for the RAID set:

- Drive order 1-2-3-4 (using $MFT file number method). Drive 3 was the partially rebuilt drive
- Stripe size 64k.
- NTFS Parameters:
* Bytes per sector 4096
* Sectors per cluster 16

I then deleted the old raidset on the physical disk and re-created a new raid+volume set. Removed disk 3, activated the raidset, and booted up. The settings seem correct because a partial scan of the volume in R-Studio was able to recover many of the files - JPEG images of ~500KB displayed properly - which should indicate the drive order and stripe size are correct. Unfortunately the old NTFS partition isn't recognized properly by WIndows. I shows the RAID volume as follows:

698.48GB Free Space
1349.51 GB Unallocated
3539.93 GB Unallocated

I'm now running a full volume scan in R-Studio to see how many of the files it can recover. I may end up doing the recovery this way, but if you have any tips on how to fix Windows discovery of the partitions on this volume, please let me know? One thing I noticed is that drives 3 and 4 for some reason had a WIndows MBR in sector 0 of the drive. For drive 4, this is probably the correct parity data (1 xor 0 xor 0 = 1), but not sure about why it would be there on drive 3. Possibly an artifact of the partial rebuild? Both of these drives were partially recognized by windows when I inserted them to create a disk image. (Had some "free space" and "RAW" parts). I did not "Initialize" them, but worried that Windows7 might have written something to the drives.

Also, do you guys know of any detailed documentation of the Areca metadata format that gets added to the beginning of an array member disk? It would be really helpful to look deeper into the images and see if Win7 has made any inappropriate changes
 
You would know if they were RAID rated drives. RAID (enterprise) drives are usually like 2-3x as expensive.

The difference is the RAID drives have a different firmware that disables TLER (time limited error recovery) and other power saving features. TLER and power saving features (head park, etc..) in consumer cheap disks can make arrays randomly loose disks by seeming to be failed.

Also enterprise HDs are said to be pre-burned in at the factory for a more reliable drive, and typically have lots more cache.

RAID 2TB drives

$224
http://www.newegg.com/Product/Product.aspx?Item=N82E16822136579

$209
http://www.newegg.com/Product/Product.aspx?Item=N82E16822148929
 
The original Seagate 2TB disks were not rated for RAID. It's pretty obvious now that a random timeout caused drive 3 to drop and start rebuilding. HW wise the drive is fine and I can still read from it (made full disk image).

The new drives I purchased are DeskStar 7K4000 which specifically state support for TLER even for the desktop version (although it looks like it gets reset to off on reboot, so will need a startup script to enable it once these are put into service).

The irony is, I've always had the best luck with Seagate drives in my desktop PCs. Started with Hitachi (failed), then WD (failed), then Seagate (was fine until the RAID - 4x750gb ran fine for years, 4x2tb failed within ~ 1 year - and no these were never burned in, so that could have been the issue as well - the drive failed in the middle of a long video transcode batch, lots of writing to new sectors... Now I know to do that haha..).

Now I'm back full circle to the DeathStars....
 
Hi all,

So, the delete/add method seems to have worked after I found the correct settings. I can see 90%+ of my files, and video files 3GB+ play without issue.

There are still some folders that don't open ("corrupted") - including My Pictures, which is the one I need to recover the most. When I try to list the contents of the drive in R-Studio, it gives me correct directory listings for these "corrupted" folders - but the files inside seem to be jumbled up (JPEGs with various strange lines and color blocks).

Another interesting piece of data - RStudio reports that $MFT is around 1.1GB, but $MFTMirr is only 65536 byes (looks like it might have been wiped?)

At this point, what's my best bet to recover these files?
1) Run a chkdisk?
2) Full scan in R-Studio?

Since I have enough spare drives, I'm leaning towards doing #2 first anyway, to avoid potentially losing any files while running chkdisk.

Does that sound like a good plan? Any other tips for fixing the file system corruption?

Thanks again all for your help!
-Dmitri
 
No chkdsk, unless you want to make the situation worse. Approach a recovery as a read only operation, avoid anything that will make changes
 
^ What I figured as well.. I'm waiting for more advice before I do anything stupid

One thing I noticed is that most of the older files are 100% OK, while most of the newer ones from the past 6 months or so seem screwed up. Not sure if that helps at all
 
Did you not have a backup? RAID is nice, but as others have state a meeeelion times, RAID is not a backup.
 
Optimally, the first step would be to make sector level backups of all drives before you do anything else if the data is important, but that would require 8 TB of free space somewhere in your case. With those backups you can always start from scratch if anything goes wrong.
 
Actually, I do have images of all the old drives created with R-Studio. (each on its own 4TB HDD). There is also free space on those drives I could use for recovering the old data via R-Studio or another software.

Yes, I learned my lesson on RAID vs backup. I do have some older copies of many files, but it looks like the older ones on the RAID are fine as well. It's mostly the recent files that seem to be corrupted.

If anyone has suggestions on other software to try, or tips for working with R-Studio, please share! Thanks!
 
You must log in or register to reply here.
Back
Top