read recover failed optical disk bad block hardware uncorrectable ecc error
Q: I cannot read or boot my optical disk anymore. In the console I see a series of READ RECOVER messages followed by a READ FAILED message.
Q: I have a bad spot on my optical disk. I tried rewriting over it, then tried erasing the whole disk, but I get an I/O error always at the same spot. Is there a way to mark the bad spot?
A: A sudden power off or interrupt can cause the erasure of a half track on the OD. One of the symptoms is a series of READ RECOVER messages followed by a READ FAILED message. It might look like:
This means that the read was to logical block 160120, physical block 161992, track 14273 of drive 0, at sector 8.
The only way to recover is to rewrite the sector, which may or may not be difficult, depending on which file (if any) the sector is located in. Try this procedure.
Fixing an optical with "Uncorrectable ECC" errors
1. Go to a machine which is not booted off of optical
2. Insert the optical.
3. Become root
4. Attempt to fsck the a partition of the disk.
5. Carefully record the physical address of the bad block
6. Run the disk program on the device /dev/rod0a
7. Run the command "abort" in disk to turn of abort on error
8. Run the "read" command on the physical sector
a. enter the physical block number
b. enter 1 for the number of sectors per transfer
c. enter 8 for the number of transfers
d. enter 1 for the sector increment
e. record the number of bad sectors
9. Run the "write" command on the physical sector
a. enter the physical block number for starting block
b. enter 1 for the number of sectors per transfer
c. enter the number of bad sectors from step "e" above for the number of sectors to transfer
d. enter 1 for the sector increment
e. enter 0 for the random data
10. Repeat step 8 above; you should receive no error
11. Rerun fsck on the a partition
a. it may be necessary to run with an alternate super-block
b. use -b 16 for this case (additional backups are in NextAnswers, os.293)
c. repeat fsck until no errors occur
12. Mount and examine the file system to confirm that it is still intact
Note also the 2.0 disk command for repairing bad ODs.