sqlite error: disk I/O error

[Michael_google gmail_Gersten]

Oh, this isn't nice.

smartctl -x disk0

...
Error 5 occurred at disk power-on lifetime: 9454 hours (393 days + 22 hours)
  When the command that caused the error occurred, the device was
active or idle.

  After command completion occurred, registers were:
  ER ST SC SN CL CH DH
  -- -- -- -- -- -- --
  40 59 07 59 16 be 40  Error: UNC 7 sectors at LBA = 0x00be1659 = 12457561
...

So it looks like I've got an error somewhere around that block. And,
other errors at LBA's 293167 and 314548. (And, I'm at around 11,000
total poweron hours.)

First: Isn't that excessively early? What are the spec'd error rates
for this drive?
Model Family:     Fujitsu MHV series
Device Model:     FUJITSU MHV2060ATPL
Serial Number:    NSA1T5825GFR
Firmware Version: 008100A0

Second: How can I force a test read or write of those sectors? How can
I tell if they are allocated? (Normal HFS+, but multiple partitions.)

Looks like fun ... but the drive's self test has never been run. Glad
I have an external TM backup before I test it :-).

(Hmm, I saved my 10.5 install disk as a partition on another machine
on the network. If worst comes to worst, can I boot that in target
disk mode and reinstall over the network? Do self tests do read-only
testing, or write tests? Obviously, a write test will be more
accurate, but a failure ...)