Fujitsu MAN3735 Manuale Utente

Pagina di 302
C141-E124-01EN
1 - 31
(3)
Operation if formatting is not completed normally
While the FORMAT UNIT command is not being executed after a change in data format related
parameters by the MODE SELECT  command, if the FORMAT UNIT command is terminated
abnormally for any reason (example: power cut off or RESET condition), the data on the disk
media cannot be accessed normally.  In this case, the IDD reports a CHECK CONDITION status
for the command accessing the disk media.
The sense data generated at this time indicate a MEDIUM ERROR [=3] / Medium format
corrupted [=31-00] in the case of the former and MEDIUM ERROR [=3] / FORMAT command
failed [=31-01].  In either case, the IDD is restored to the ready state by reinitializing the disk
media using the FORMAT UNIT command.
1.7.5
Error recovery processing
If an error is detected an recoverable error on the SCSI bus or internally in the IDD, the IDD
attempts recovery processing of the detected error during command execution.
The INIT can specify detailed parameters related to error recovery processing and whether or not
there is a report for errors which were successfully recovered by using the MODE SELECT or
MODE SELECT EXTENDED command.  It can also save those parameters in accordance with
INIT specifications in the system space on the disk media, and can change them temporarily
without saving them.  After the power is switched on, or after a reset (RESET condition or
TARGET RESET message), the saved parameters are read and the error recovery processing
procedures is initialized.
See Section 3.1.5 “MODE SELECT” and Section 3.1.6  “MODE SELECT EXTENDED” for details of
the MODE SELECT or MODE SELECT EXTENDED command.  See Section 5.3 “Disk Drive Error
Recovery Processing” and Chapter 3, “SCSI Bus Error Recovery Processing” of “SCSI Physical
Interface Specifications” for details of error recovery processing.
An outline of error recovery processing executed by the IDD is shown below.
(1)
Recovery processing for errors on SCSI bus
The IDD does not perform recovery processing for error detected on the SCSI bus.  When error
recovery is impossible, or when a critical error related to SCSI bus protocols is detected, the IDD
may transfer the SCSI bus forcibly to the BUS FREE phase and clear the command that is
currently being executed.
Table 1.4
Outline of SCSI Bus Error Recovery Processing
Item
Type of Error
Error Recovery Processing
1
MESSAGE OUT phase parity error
“CHECK CONDITION” Status
2
COMMAND phase parity error
“CHECK CONDITION” Status
3
DATA OUT phase parity error
“CHECK CONDITION” Status
4
Received INITIATOR DETECTED ERROR message.
“CHECK CONDITION” Status
5
Received MESSAGE PARITY ERROR message.
“CHECK CONDITION” Status
6
RESELECTION phase time-out
Re-execute RESELECTION phase.