HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 77
HP restricted
7 Exception handling
These pages cover methods of dealing with certain error conditions and exceptional circumstances.
They include the following:
• A suggested Escalation Procedure to follow when exceptions occur
• How to make the most of the TapeAlert facility, see “Supporting TapeAlert” on page 78
• How to respond to the ‘Clean’ LED (lit on the front panel when there are an excessive number of
retries or error corrections)
• The need for a pass-through mode, which can return information or commands specific to a
vendor’s product
• How drives recover from read and write errors
• How to use the drive’s ability to read through media errors, so that as much data can be
retrieved from a badly damaged tape as possible
Typical escalation procedure
For exception handling, there needs to be a well defined escalation path, through which the calling
application, user, operator or System Supervisor may take increasingly drastic action to clear any
product-related faults.
An escalation procedure is important to allow local recovery where possible, and to avoid the
unnecessary replacement of peripheral devices.
A typical escalation procedure is as follows:
1.
Retrieve fault information.
You can run HP’s Library and Tape Tools (L&TT) and obtain a ticket, or you can retrieve specific
fault information from the following sources:
•I
NQUIRY data, such as firmware revisions
•R
EQUEST SENSE data, such as Additional Sense Codes and Drive Error Codes
•M
ODE SENSE data, such as data on the current configuration
•L
OG SENSE data, stored in the drive’s logs
2.
Inform the user.
The system gives the user helpful advice by attempting to decode the returned information, and
also allows the user access to the raw data.
3.
Allow the user to try recovery.
For any fault, the system allows the user to use simple recovery commands such as
REWIND or
LOAD/UNLOAD.
4.
Allow the user to reset devices.
If these actions fail, including repeated attempts to retry the operation, the user should be able to
reset devices on the SCSI bus selectively, through the use of a
LOGICAL UNIT RESET Task
Management function. This function is specific only to the device to which it is addressed and
will not disturb other devices on the SCSI bus.
