Hey guys, had an old 1.5TB drive that started showing errors during parity checks so I swapped it out for a new 3TB disk.  I ran two preclears on the old 1.5TB drive and it has around 600 sectors pending re-allocations at the end of the preclear (both times).  What can I do with it?  Is it a lost cause?  The drive has no warranty.

 

What do you guys usually do with drives in this state?  Use it as a paperweight?  Or how do you dispose of it?

 

Thanks

Hit it with a hammer and send it to platter heaven.

 

Kryspy

In my experience, some drives develop these pending reallocations that never clear. But at the same time they also never generate true reallocations or generate read errors either. I am not sure what this means, maybe a firmware bug or maybe a legitimate condition.

 

Some on the forums suggest using these types of drives to store data backups. With unRAID and its protection, the chances of the disk being called into service to cough up its data is small, and if this is a part of a parity protected backup array, the chances of it being available in the unlikely event you should need it are very good.

 

That would be my suggestion.

Considering it's only a 1.5TB disk and more than 3 years old (since it's out of warranty) I'd personally scrap it rather than risk putting data on it, but that is just me. Since you can pick up a 2TB disk for less than $100, why risk it?

Backup data on it...reformat it...and store 'non-critical', but 'what a pain to recreate' data on it.

You wont trust the drive with your data now. Are you willing to trust the drive with your data later?

 

You can try doing a multipass badblocks in write mode to see what it does.

Still, I'm not sure I would use it.

 

Did you do a smart LONG test? That gets the firmware into action.

 

You can try a smart long test. See if there are read errors at LBA,

do a multipass badblock in write mode and another smart LONG test.

 

If the last smart LONG test still shows pending sectors or read error at LBA, scrap it.

 

fwiw,

In days of yesteryear, drives always had bad sectors.

badblocks would weed them out and document them. Then you would pass that file to the mkfs which would mark them bad in the filesystem data, thus avoiding them.

 

If you really want to use the drive for some kind of archival backup you could do that.

I believe NTFS can do the scanning also.

... started showing errors during parity checks ...

 

This drive is showing errors ... so I'd simply get rid of it.    If you were simply upgrading your array to larger drives, I'd suggest using it for your backups ... static backups aren't likely to be accessed very much, and if you store MD5's for all the files on it, you can also easily validate that they're still good at any time (in fact, it's a good idea to do this with your backup disks at least once/year).

 

You could, if you want, do a full Level 4 pass with Spinrite (this will mark any bad sectors so they're not used); and then format the drive to use for backup purposes ... but for any drive that has ANY errors, I agree with bkastner -- new drives are very inexpensive ... so I'd just toss the drive.

 

Back in the MFM days, I found it extremely satisfying to open up the bad hard drive to reveal the platters, then power it up and press the drive head onto the platter to make pretty designs.

 

I mounted another platter from the drive in a picture frame and hung it at my desk.

Back in the MFM days, I found it extremely satisfying to open up the bad hard drive to reveal the platters, then power it up and press the drive head onto the platter to make pretty designs.

 

I mounted another platter from the drive in a picture frame and hung it at my desk.

Coasters?

I am about as brutal with decisions on kicking a drive as anyone.

 

But I've seen several of these drives that don't reallocate anything and have a bunch of pending sectors that don't change.

 

Not saying I'd love it in my primary array, but I'd use it in a backup all day long.

 

If you're going to toss it, feel free to ship it my way. It wouldn't go to waste.

I have a 2TB drive with 406 reallocated sectors. That's way to many to use in an array, so I use it as my cache drive, but that cache drive is only used as a "temp" drive for nzbget and for data I don't care about. nzbget unpacks from the cache drive to an array drive. When unpacking, unrar will do a CRC check, so I don't worry about data integrity on the cache drive.

 

It had 405 reallocated sectors and that went up by one over x months. I keep an eye on it and if it keeps increasing, I'll probably ditch it.