Should I replace the HD?

Question

as mentioned in a former topic, I had a severe crash after installing SL.
Everybody advised me to change the HD, although every tool I took proved it to be o.k....
Since I reinstalled Leopard (not SL), everything seems to be fine, except using the migration assistant, which does not seem to work, so I could not prove if my usual stuff (i.e. Logic studio/CS4) runs properly.
Should I replace the HD? Does the HD still cause problems?
Any help?

Answer 1

I don't disagree with most of what you say but a few things might be worth considering.
AFAIK, all modern ATA drives use some form of low level error correction & detection, typically CRC, & automatically map out blocks that fail that test after some number of retries. This occurs on all normal reads of physical blocks & is independent of the host computer.
Utilities like TTP at one point (& possibly still today) supplemented this by performing tests that in effect artificially reduce the number of retries to a very low number on the theory that the drive will allow "weak" blocks (those that take several retries to pass error correction tests) to continue to be used until they fail, causing data loss. These "weak" blocks would then be mapped out in some way by the utility. IMO, the value of this is debatable because in effect it assumes that a third party vendor knows more about preventing data loss than the manufacturer of the drive. It also can map out blocks that have no real problems since this kind of "soft" error (failed CRC on a single read) occurs routinely in modern very high density drives.
Absent this kind of utility overruling the disks own checks & mapping out what might be perfectly good blocks, it is sometimes useful to check how many spare blocks have been mapped in & just as importantly how many remain because this tells you something about a suspect drive's continuing reliability. This is where I see vendor utilities and/or SMART Utility for the Mac being of some value. However, since most drives' built-in SMART systems monitor the same thing, I don't think most users need to worry about this … at least for internal drives for which SMART monitoring is supported by the OS.
BTW, by "increasing frequency" of bad blocks, what I meant was something I have observed in the past: when the bad block count suddenly starts growing, it is usually a sign of impending drive failure. Usually this is because the drive's platter has shed magnetic media, perhaps because of a head crash, & the debris are getting trapped between the head & platter & gouging out still more media. Again, if SMART monitoring is available for the drive, it should detect this.