I recently switched from a Norco 4020 to a Norco 4224 case. I had to buy 6x SATA to SAS to connect to the new case, and another SATA controller to fully connect all 24 hotswap bays. Everything went great, parity sync was fast with 0 errors. 19 Data drives, and a Parity drive. However, about a week later after transfering about 150GB to it, disk12 showed 36 errors on it. I did a SMART test and it shows Current_Pending_Sectors = 1422, Offline_Uncorrectable = 1084, and Multi_Zone_Error_Rate = 8032. This is 100% the drive, correct? Or is it possible I have something going on with a cable, the case, or a SATA controller?

 

Syslog was to big, so I copied just the part where I see errors.

 

SMART: http://img849.imageshack.us/img849/5616/errors.png

 

EDIT: Also should I be worried about the 150GB I just transferred? Possible that it got corrupted?

EDITX2: Just confirmed the content I transferred IS corrupt and skips randomly, luckily I have a backup.

 

syslog.txt

Well, i'm pretty sure its the disk.

 

If I try to transfer my data off the drive I get read errors. I'm not sure what I should do, i've set up an advanced RMA to get the drive replaced.. but what am I suppose to do while that gets here? If I transfer the data off the drive, theres a good chance the read errors will cause the data to be corrupted? So I should wait for the new drive, remove this drive, then have unraid rebuild? Or transfer my data off now?

Well, i'm pretty sure its the disk.

 

If I try to transfer my data off the drive I get read errors. I'm not sure what I should do, i've set up an advanced RMA to get the drive replaced.. but what am I suppose to do while that gets here? If I transfer the data off the drive, theres a good chance the read errors will cause the data to be corrupted? So I should wait for the new drive, remove this drive, then have unraid rebuild? Or transfer my data off now?

It you un-assign the drive anything read from it will be re-constructed from parity and the other disks.

Well, i'm pretty sure its the disk.

 

If I try to transfer my data off the drive I get read errors. I'm not sure what I should do, i've set up an advanced RMA to get the drive replaced.. but what am I suppose to do while that gets here? If I transfer the data off the drive, theres a good chance the read errors will cause the data to be corrupted? So I should wait for the new drive, remove this drive, then have unraid rebuild? Or transfer my data off now?

It you un-assign the drive anything read from it will be re-constructed from parity and the other disks.

 

So the best method would be to just wait for my replacement drive, preclear the new drive, then unassign the bad drive and reassign the new precleared drive and let parity rebuild it?

 

If this is the case I think I will just overnight another 2TB green (currently $83.98 after shipping with prime), to reduce my chances of another drive going bad while I wait for western digital.. then I can just use the western digital drive as a back-up drive incase this happens again.

Sounds like a solid plan.

Have you done a parity check? If you do a parity check and the drive reads wrong then it could mess up parity. Otherwise, you are good to go with a replacement. unRAID would have properly created the parity when you wrote to the drive even if the drive didn't save the data correctly. If you just power down and pull the drive you can start the array without it. Once started, the drive will be simulated and you can copy the contents off it.

 

Peter