In addition, the application can determine which TapeAlert flags are supported by the device server through the TapeAlert supported flags VPD page (B2h).

TapeAlert polling usage model

In this model, the application configures the device server by setting the TASER bit in the Device Configuration Extension mode page (10h) to one.

The device server does not notify the application that a TapeAlert flag has changed. The application can read the TapeAlert log page or the TapeAlert Response log page at any time (for example, polling every x seconds) or upon certain relevant operations:

Before loading a tape

Immediately after detecting an unrecoverable error

Before unloading a tape

At the end of a data transfer

TapeAlert informational exception usage model

In this model, the application configures the device server by setting the TASER bit in the Device Configuration Extension mode page (10h) to zero, and in the Informational Exceptions mode page, by setting the DExcpt bit to zero and the TEST bit to zero.

The device server notifies the application that a TapeAlert flag has been activated by reporting and informational exception condition on the next SCSI command, with additional sense of 5D00h (failure prediction threshold exceeded). If the device server has been configured to return descriptor format sense data, the current state of all TapeAlert flags appears in the Information sense data descriptor. If not, the drive returns fixed format sense data and the application should read one of the TapeAlert log pages to retrieve the state of the TapeAlert flags.

For more information see SSC-3 rev 02, section 4.2.15.

Reading the TapeAlert log

Each time the TapeAlert log page is read, the application should follow this procedure:

1.Read all 64 flags to discover which are set (there may be more than one).

There may also be data associated with a set flag in the remainder of the flag byte, which should also be read.

2.For each flag that is set, log the associated error message.

3.Notify the user through the normal methods (such as broadcast, E-Mail, SNMP) by displaying the error message suggested in the table. Include the following:

Details to identify which drive the message refers to.

The software label of the tape cartridge when relevant.

The severity of the message (Information, Warning or Critical, with Critical the most severe). Where there are several flags set, list the messages in ascending order of severity.

4.Apply any error message filters in the software to the TapeAlert errors. If several TapeAlert flags are set, they should if possible be presented to the user as a single event. For example, the error messages could be displayed together in a single message box.

88 Exception handling

HP restricted

Page 88
Image 88
HP LTO 4 SCSI Reading the TapeAlert log, TapeAlert polling usage model, TapeAlert informational exception usage model