On May 4, 2005, at 9:49 AM, Ari Sovijarvi wrote:
> I'd find it very odd, if BOTH of the drives were suffering the same 
> kind of a problem, especially when they're about 3 months old. Then 
> again, this is a RAID system, so I could replace the drives one by 
> one, but I'd rather look for other alternatives first before spending 
> 150e for new hardware that might not fix the problem.

Give a thought to testing the drive either with smartmon or offline 
using the manufacturer's test utility.  (I don't know for sure that 
NetBSD supports the needed ATA capabilities, but if pkgsrc has 
sysutils/smartmontools, then it probably can.)

Running a long SMART self-test will do a good job of examining the 
drive hardware and ought to give you a better picture of what is going 
on.  It would not surprise me if this slow drive has a bunch of bad 
sectors which are being transparently corrected using ECC and the 
automatic hotspare mechanisms that modern drives do.  Even though the 
drive hasn't reported a failure, the correction process would cause a 
very noticable slowdown, similar to what you've described is actually 
happening.

-- 
-Chuck