so today while copying some files to a new drive, my disk2 was taken offline and showed over 200 errors.

thinking it might have been my fault (wiggling a cable loose or whatnot when i installed the new disk), i unplugged and plugged in all sata cables to ensure they're seated correctly and followed the directions to make unraid trust my disk. unfortunately, it came back with over 300 errors.

i ran a short smart test (included in attachments) and figured i'd unassign the disk and run a pre-clear before deciding to RMA.

just want an expert's opinion on whether or not i should rma it.

i've also included my syslog --- i reboot the system after unassigning the drive and put it in a new onboard sata slot to pre-clear so this log is while the pre-clear is going (read state).

syslog-2012-03-23.txt

smart-1015pm.txt

If you caused the errors by wiggling a cable there is nothing wrong with the drive.  The SMART report shows no re-allocated sectors.

Start the array with the disk un-assigned.  That will cause the aarray to forge its model/serial number.

Then stop the array, re-assign the drive to its slot, and re-start the array.

unRAID will re-construct the data to the drive.

 

When the re-construction is complete you'll be back in full working order.

 

Joe L.

here we are 12 hours later.

syslog is throwing a bunch of media/unc errors now that the preclear is writing to the drive.

ran another smart report and it doesn't show anything different than before. what gives?

my gut is saying to rma this 2 week old drive (will never get a WD drive again).

 

edit: another quick question:

i understand that the parity will rebuild the disk when a new one is put in.

i also understand that the parity is currently not protected, so in the event of another disk failure, i'd lose data on 2 disks.

however, what i'd like to know is if it is advisable to shutdown the server until i put a new disk in place of the failed one?

syslog-2012-03-24.zip

smart-1015am.txt

Basic answer: yes. Shutting down the system and waiting for a new drive is a preferred thing to do if you want to lower your risk profile.

 

Also consider getting an extra drive, add it to the system, preclear it but do not add it to the array, that way if something like this happens again (and it will; drives fail) then you will be up and running without downtime.. (and the disk can be kept spun down awaiting failure)

Basic answer: yes. Shutting down the system and waiting for a new drive is a preferred thing to do if you want to lower your risk profile.

 

Also consider getting an extra drive, add it to the system, preclear it but do not add it to the array, that way if something like this happens again (and it will; drives fail) then you will be up and running without downtime.. (and the disk can be kept spun down awaiting failure)

ok, system is going to be shut down until i replace the drive.

couldn't have happened at a worse time. just had a samsung drive fail and i get the replacement on monday.

and now, lucky for me, i get to go through the rma process with WD.

guess on the brigher side, when i get my rma back from WD i'll have it precleared and ready for the next failure.

just a quick follow-up and for anyone reading.

i pulled the disk from my unraid server and put it in my windows 7 machine.

ran hdtune on it and reported the same as the unraid smart report (1 pending sector).

ran a few read/write tests on it with hdtune and cleared up the pending sector.

wrote 0's to the drive with hdtune, no errors.

popped the drive back into my unraid server and ran a successful preclear cycle on it.

adding it to the array now, hopefully the rebuild goes ok and i'm back in business.

 

edit: eh, wishful thinking i guess.

syslog started throwing a bunch of errors regarding the same disk (disk2) when i started the rebuild.

sending out monday for rma.