All hard drives have some bad blocks. Not only that but they also have spare blocks available as well. The idea is when the drive attempts to write to a bad block it verifies the write and if there is still an error after a given number of retries the drive itself flags the data block as bad and remaps that location to one of the spare blocks that is placed on the drive for just that purpose. Unfortunately blocks can go bad after there is already data written there. The TechTool Pro surface scan detects these bad blocks and in some cases is able to remap them on the fly, but if there is data written on the block it may have already been corrupted.
The only permanent fix for this condition is to backup everything you can, then erase the drive using either zero all data or 8 way write option to force remapping the bad blocks. This works as long as there are remaining spare data blocks. Once you run out of spare data blocks or data blocks start going bad quite rapidly this
should show up in the S.M.A.R.T. values although there are no guarantees.
TechTool Pro detected a similar situation on my wife's G4/733 that had a so called
Deathstar drive in it. I solved the situation by:
- Purchasing and installing a second hard drive that was quite a bit larger than the original
- Used Carbon Copy Cloner to clone the bad drive to the new drive
- Booted from the newly cloned drive
- Erased the original drive using the 8 way write option
- Tested the old drive with TechTool Pro and it got a clean bill of health.
The old drive is now used for backups and has perked along happily for over a year now.
If you decide to go this way, I would switch the old drive from the ATA100 bus to the ATA66 bus and put the new drive on the ATA100. That avoids bus conflicts and optimizes drive I/O. The jumpers on your new drive should be set to
cable select and not Master or slave. The jumpers on your existing drive are already
cable select and you can leave them alone. This is true even if you elect to put both drives on the same bus.
Let us know the results of rerunning the TechTool Memory Test. If that fails again there may be an underlying reason that both the HD and RAM would go bad.