[mdlug] OT: Hardrive - Good or Bad

[Paul]

Stan Green <Stan at mcomputersolutions.com>, on Mon Jul 27, 2009 [05:35:26 PM] said:
> I had a EIDE 250G WD drive in a RAID 5 array. The system failed the drive, so 
> I replaced it. (Yes, I am 100% sure I took out the right drive.) So a few 
> months later, I decided to see I the drive just had bad blocks and could be 
> fixed. So I put the drive in another PC and ran mke2fs -j -cc. This ran fine 
> and reported no errors. Then I ran smartctl -d ata -t long. This also 
> reported  (smartctl -H) no errors.
> 
> So what am I to believe; RAID, mke2fs or smartcl to tell me if the drive is 
> good or bad? Given 2 out of 3 say it is good, should I trust it?
> 
> Thanks,
> Stan Green

	Hi;

	The drive may have had an unrecoverable read error on one
or more sectors. The kernel logs would note this.
	Later, writing to the bad sector could cause the drive
to remap that sector to a good one. Looking at smartctl -a 
should show if this happened. A drive will also sense a sector is
starting to go bad, and remap it before its unreadable. You can
see if this is happening by examining the smartctl -a numbers.
Lots of remapped sectors, or an acceleration in such, could
indicate a drive that is heading toward failure.
	Otherwise, it could have been a cable problem, memory
corruption, or a controler bug.

Paul
set at pobox.com