Sign in with
Sign up | Sign in
Your question

Raid 0 barely functioning - Can't tell if faulty drive or other issue

Last response: in Storage
Share
September 10, 2012 10:38:30 AM

Mother Board: GA-X58A-UD7 (rev. 2.0)
Raid: 0 (Stripe)
Hard Drives: 2x Hitachi Deskstar 7K3000 3TB HDS723030ALA640
Array Size: 5589.0000GB
Block Size: 128 KB
Controller: JMicron JMB362 (Gigabyte SATA2 Sata Controller)
Array up time: about 12 months

- Both drives are visible and accounted for in start up
- Drives are used for main storage (Not OS / but linked to desktop)
- Windows boots fine & Raid drive is accessible
- Accessing files on desktop is fine
- Accessing files in large directories causes freezes and long periods of loading
- Copying files to another drive always start normally, but quickly drop to about 213 KB/second
- The larger the folder and the greater the number of files, the bigger the struggle
- Explorer constantly crashing while attempting to copy files

My theory:

- One of the drives is failing and is on its last legs. Probably due to mechanical fault.

At the moment I'm proceding with the very stressful task of backing up some important work documents since I've been complacent with backups lately (Last backup May 25th)

:sweat: 

But the biggest disappointment of all is losing over 3 Months of vinyl rips, that's about 125 GB of files... Which at the moment seems impossible to back up without complete stalling and eventual crashing.

I'm pretty desperate for any help or suggestions, anything will be much appreciated.
a c 312 G Storage
September 10, 2012 7:33:44 PM

I would try to retrieve the SMART reports from each drive. Look for reallocated, pending, or uncorrectable sectors.

Then use a cloning tool such as ddrescue to make a sector-by-sector image of the failing drive.
September 10, 2012 9:47:17 PM

fzabkar said:
I would try to retrieve the SMART reports from each drive. Look for reallocated, pending, or uncorrectable sectors.

Then use a cloning tool such as ddrescue to make a sector-by-sector image of the failing drive.


Thanks, that sounds great.

So basically I would need to temporarily break the raid, analyse both drives for dodgy sectors, figure out which drive is the culprit, then use a cloning tool to make a sector-by-sector image of the failing drive, copy that image on to a new but exact same model drive and re-enable the raid array?

- It seems safe enough but aren't there risks of destroying the array in the process?

+ I'm not sure I need to analyse the drives, after closer inspection I've discovered that one of the drives seems to be suffering the click of death
a c 312 G Storage
September 10, 2012 10:24:35 PM

I don't know what risks there would be in breaking the RAID. That question would best be anwered by someone else. However, you might be able to clone the entire RAID in situ, but that would require a 6TB destination array.

It appears that you have already uncovered the culprit, but you may have been able to retrieve the drives' individual SMART reports using smartctl (from smartmontools). Smartctl was written for Linux, but it has been ported to Windows.

The following two utilities provide Windows GUIs for smartctl:
http://gsmartcontrol.berlios.de/home/index.php/en/Home
http://code.google.com/p/hddguardian/

Normally you can't see each individual HDD when it is in a RAID, but smartctl can at least work with some RAID controllers.
!