Auto Alternate Block Allocation Processing - Fujitsu MAX3073NC - SCSI,80, Dell DC961,Hard Drive Scsi Logical Interface Specifications

Sense Data Error Recovery Methods
If error correction is permitted but immediate correction is not permitted (DCR = 0, EER = 0), reading of
data as in item (2) is retried up to the number of times specified in the other MODE SELECT parameters
("Number of Retries during Read" or "Number of Retries during Write").
If error correction is prohibited (DCR = 1, EER = 0), reading of data as in item (2) is retried up to the number of
times specified in "Number of Retries during Read" or "Number of Retries during Write."
If the error cannot be recovered from even when this retry processing is executed, the IDD terminates the
command that is currently being executed with a CHECK CONDITION status. The sense key in the
sense data at this time is "MEDIUM ERROR [=3]" and the additional sense code indicates "Unrecovered
read error [=11-00]."
(4) Other internal IDD errors
If an irrecoverable error other than those in items (1) to (3) above is detected internally in the IDD, the IDD
terminates the command that is currently being executed with a CHECK CONDITION status. The sense key
in the sense data at this time is "HARDWARE ERROR [=4]."
5.3.2

Auto alternate block allocation processing

The IDD is equipped with a function which automatically executes alternate block allocation processing
for defective data blocks detected during execution of the READ and READ EXTENDED commands.
(1) Auto alternate block allocation processing during a read operation
This function is enabled by specifying "1" in the ARRE flag in the MODE SELECT parameters.
Furthermore, this function is applicable to the following commands only.
• READ command
• READ EXTENDED command
When this function is permitted, if, during read processing to the data area, an error which is an object of retry
processing is detected, and which it is impossible to recover from even when reading is retried (see Section
5.3.1 (2)) the number of times specified in the "Number of Retries during Read" field in the MODE SELECT
parameters, the IDD allocates an alternate block for that data block and executes copying of the alternate block
to the data area. Alternate block allocation processing is the same as processing with the REASSIGN
BLOCKS command. Furthermore, recovered data copied to the alternate block are data which are read out,
skipping the ID area, if there is an ID area, or they are data which have been corrected through the ECC.
This function is applied twice only during execution of one command. Alternate block allocation
processing and copying are executed for the first and second defective data block to be detected during
execution of a command, but if a third defective data block is detected, the command that is being
executed at that point is terminated with an error. However, by the INIT's reissuing the same command,
auto alternate block allocation is applied sequentially and error recovery can be attempted.
5-24 C141-C010