I have run UnRaid for a long time (V3), and whenever I have had a drive redball, I have simply replaced it without question. To date drives which have failed have always been either very near to the end of their warranty period (In which case I have always got a new drive under RMA), or well outside of the warranty period.

 

I have recently added a 2TB drive to the array, which had previously used for nightly backups, it is almost 2 years old but has only 91 power on hours. Within a month this drive has redballed?

 

I have checked all power and data connections (drives are mounted in 5x3 enclosures), but drive is still redballed.

 

So I installed a new drive, data reconstruction has been carried out and the server is now Ok

 

I have performed SMART tests and all diagnostic tests using WD Tools, and the failed drive tests out OK? The drive has been fully zeroed, CHKDSK has been run and the drive fully formatted twice. I now intend to carry out a number of preclearing cycles too.

 

I now don't know why the drive redballed, of if I should trust it in the future. I am looking to use the drive to replace an old 1TB drive in the array but am not sure.

 

I would welcome comments and advice.

Yours is a great story about the value of unRaid in maintaining a healthy array over a long period!

 

About this drive, can you comment on what you were doing with it when it red balled? Were you writing to it? Doing a parity check? Doing nothing at all?

 

I have found the greatest wear and tear on a drive is sitting for long periods. Like a car parked and not driven.  I have a stack of IDE drives that refuse to spin up that were fine when taken out of service.

 

It seems like you've stressed the drive pretty hard since the problem happened and it's working well. Red balls are usually caused by cabling issues and maybe you inadvertently fixed it. I'd probably give it another chance, but do parity checks a little more frequently than normal.

Thanks for the reply, the drive was spun down when it was redballed.

 

The drive was OK one day, I woke up the next day, the whole array was still spun down but there was a red ball shown against it? I tried reseating power & data connectors a few times, and tried rebooting a number of times, but could not get the red ball to go.

 

I appreciate that all drives have a finite life, but this drive was barely used (91 hours over nearly 2 years) and had been used faultlessly. However in a month it had reballed?

 

Even more confusing is that having tested it, including WD Tools, it proves and tests out as OK? So why did it redball in the first place. With hindsight I should have kept a syslog, but it's too late now.

 

As suggested I am tempted to reuse it, but will subject it to a number of preclear cycles to stress test it first.

 

Once a drive is red balled, nothing you do will make the red ball go away short of rebuilding the disk onto itself or some other disk or using the trust my array procedure.

Correct.  Once a disk has had a failed "write" to it, it will never go back into service by simply re-seating the cables/connectors, even if they were the cause of the write failure. 

It will stay"disabled" (red)

 

until...

 

It must either be re-constructed onto itself, or another drive, or you must force it back into the array by setting an entirely new disk configuration.

See the wiki for details.

Side note on this "only a write triggers a red ball" ...

 

A read error will result in unRAID trying to reconstruct the sector and will issue a write to the offending drive, and if that write fails the drive is red balled. This is the likely scenario of an unresponsive drive (e.g., normally caused by bad cabling or power). So it really doesn't matter what the user was doing (reading, writing or a parity check), if a drive is acting up unRAID is going to write to that disk and if it fails, kick it from the array.