HP LTO 4 SAS, LTO 4 SCSI, LTO 4 FC User Manual

Download

HP LTO Ultrium tape drives technical reference manual

LTO 4 FC, SCSI and SAS drives

volume 2: software integration

Edition 1, June 2007

HP restricted

Legal and notice information

Hewlett-Packard Company makes no warranty of any kind with regard to this material, including, but not limited to, the implied warranties of merchantability and fitness for a particular purpose. Hewlett-Packard shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material.

This document contains proprietary information, which is protected by copyright. No part of this document may be photocopied, reproduced, or translated into another language without the prior written consent of Hewlett-Packard. The information is provided “as is” without warranty of any kind and is subject to change without notice. The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein.

Revision history

Version Edition Date Changes

LTO 4 1 June 2007 LTO 4 full-height drives

This document is frequently revised and updated. To find out if there is a later version, please ask your HP OEM Representative.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration

HP restricted

Related documents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Documents specific to HP Ultrium drives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Documentation map. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Drives—general . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Installation and configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Cartridges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Maintenance and troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Dealing with errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

LTO Ultrium features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

General documents and standardization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

1 Designing backup applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Optimizing performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Large data transfer size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Data compression control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Non-immediate commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Managing the use of tapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Information in Cartridge Memory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Cleaning tape heads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Monitoring tape use. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

TapeAlert . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Diagnostic logs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Displaying drive information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Drive tests. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Design goals for LTO backup applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2 Configuration and initialization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Operating system drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Inquiry string recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Support for additional LUN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Enabling additional LUN support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Supporting additional LUNs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3 Use of tapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

LTO cartridge memory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Identifying tape cartridge types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Tape status and capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Finding the remaining capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Interpreting Log Sense data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Using the SET CAPACITY command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 3

HP restricted

WORM media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

How WORM media works. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Changes to SCSI commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Re-writing media labels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Allow overwrite of last filemarks before the EOD data set . . . . . . . . . . . . . . . . . . . . . . . . 23

Using CM to check tape integrity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Behavior with a missing or inconsistent EOD value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Unique media identifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Barcode support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Responding to Cartridge Memory data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Load count . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

RWW retry counts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

4 Factors affecting performance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Ways of optimizing performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Detecting the drive’s speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Ensuring the recommended minimum transfer sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Media type identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Using Cartridge Memory instead of tape headers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Using the Performance Log page for diagnosing problems . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Time-out values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Recommended support of log pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Factors affecting performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Host-related factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Drive-related factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Format-related factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

5 Supporting Ultrium features. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Cartridge Memory (LTO-CM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Automation interface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Automation/Device Interface (ADI). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Automation Control Interface (ACI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Modes of usage through ACI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

ACI command set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

ACI commands that affect drive streaming performance . . . . . . . . . . . . . . . . . . . . . . . . . 38

New features in ACI 4.3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Further details. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

Supporting the ACI protocol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

Recommended ACI time-out values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

Treatment of reserved fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Recommended power-up sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Recommended load-unload configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Recommended Get Drive Status polling frequency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

ACI protocol communications retry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Upgrading the drive firmware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Handling irregular cartridges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

Frequently asked questions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

HP restricted

Cleaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Resetting drives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Further details. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Backup software. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Controlling data compression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Other mode page information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Accessing Cartridge Memory without threading the tape . . . . . . . . . . . . . . . . . . . . . . . . . 51

Buffer size at EW-EOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Synchronize at EW-EOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Write delay time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Rewind on reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Partition size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

6 Sense keys and codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

Sense keys—actions to take . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

Additional sense codes—actions to take . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

0h—NO SENSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

1h—RECOVERED ERROR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

2h—NOT READY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

3h—MEDIUM ERROR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

4h—HW ERROR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

5h—ILLEGAL REQUEST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

6h—UNIT ATTENTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

7h—DATA PROTECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

8h—BLANK CHECK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

Bh—ABORTED COMMAND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

Dh—VOLUME OVERFLOW. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

7 Exception handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

Typical escalation procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

Monitoring the condition of the drive and media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

Supporting TapeAlert . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

Flags. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

Designing software to use the TapeAlert log. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

TapeAlert models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

TapeAlert polling usage model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

TapeAlert informational exception usage model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Reading the TapeAlert log . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

One-Button Disaster Recovery (OBDR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

Supporting OBDR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

Responding to the ‘Clean’ LED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

Providing pass-through mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

Requirements for drivers and logical device managers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 5

HP restricted

1 Designing backup applications

In today’s computer market, software applications that use tape drives to copy the information from a computer’s hard disk for safe keeping are readily available for many different operating systems. Unfortunately, not all these applications take advantage of the advances made in tape technology over the past few years. This section examines some of the characteristics that a good backup utility should include.

Optimizing performance

There are some fundamental things that tape management applications should implement when dealing with Ultrium drives:

• Use large data transfer sizes.

• Control and monitor data compression.

• Ensure directory information is safe and accurate.

• Maximize the use of the tape drive’s internal buffering capability.

Each of these is discussed below.

For more information on optimizing performance, see “Factors affecting performance” on page 27.

Large data transfer size

Applications should use large data transfer sizes to make better use of the Ultrium drive’s internal buffers. A good goal to set is at least 128 KB each for read or write operation, with an ideal target of 256 KB:

• For fixed-length block mode reads and writes, provided the block size multiplied by the number

of blocks to be transferred is at least 128 KB, drives will provide peak performance. Small block sizes (512 bytes) are acceptable so long as they are written and read in fixed-length block mode using large transfers.

• For variable-length block mode reads and writes, the transfer length should be at least 256 KB.

Data compression control

Ultrium drives have built-in hardware data compression. Backup applications should incorporate features to report the actual compression ratio achieved during backup operations.

The typical compression ratio achieved during backup operations on PC and UNIX networks is 2:1, but this can vary widely depending on the actual data being compressed.

For more information, see “Controlling data compression” on page 51.

Non-immediate commands

Performance can be improved by only using immediate mode WRITE FILEMARKS commands.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 11

HP restricted

NOTE: Using immediate mode with other commands does not improve performance and can

cause problems when writing a driver. The SCSI specification requires that if a command is issued with the IMMEDIATE bit set to 0, the drive must flush its data buffer before it carries out the operation. This takes time.

Managing the use of tapes

The Ultrium format enables applications to monitor the performance of tapes closely, to indicate when tape heads need cleaning, and when a tape should be discarded.

See “Use of tapes” on page 19 for more information.

Information in Cartridge Memory

The LTO Cartridge Memory holds a number of pages of information that contain data about the tape’s history, such as the amount of data written to and read from the tape, the number of times a cartridge has been loaded and the tape threaded into a drive, and the number of read or write errors that have been encountered by drives with this tape. This information can be used to warn against backing up onto a tape of dubious quality, or one that is reaching the end of its life.

Cleaning tape heads

The ‘Clean’ LED on the front of HP Ultrium drives indicates when a cleaning cartridge should be used. There are two ways for backup applications to determine when the tape heads need cleaning and to prompt the user to clean the drive:

• Use TapeAlert—see “Monitoring the condition of the drive and media” on page 78 for details.

• Send a SCSI

the drive needs cleaning.

In an automation context, the tape drive tells the automation controller that a cleaning tape needs to be used through two bits in the ACI Get Drive Status command.

• The Cleaning Needed bit signals deterioration in the write or read margin of the drive and

indicates that a cleaning cartridge should be used as soon as possible. Once the drive has been cleaned successfully, the Cleaning Needed bit will be cleared.

• The Cleaning Required bit indicates that the drive is unable to read or write unless the drive is

first cleaned, so a cleaning cartridge should be used immediately. Following a successful clean, the Cleaning Required bit will be cleared.

REQUEST SENSE command to look at the CLN bit in the sense data. If the bit is set,

Monitoring tape use

Drives can report the actual amount of data that has been written to the tape, and the amount of available space on the tape. From this information, applications for Ultrium drives can be designed to calculate the percentage of tape used, and give the user feedback on the actual progress of the backup operation. This is a significant improvement over other technologies, such as DC6000 QIC products, that require the application to estimate what is going on.

Designing backup applications12

HP restricted

See “Tape Capacity Log Page” under the LOG SENSE command in Chapter 3 of The SCSI Interface, Volume 3 of the HP LTO Ultrium Technical Reference Manual for more information.

While the reliability of tape products and applications is getting better all the time, problems do still occur. There are some very simple techniques that could be incorporated by application developers to simplify the process that a user must go through to resolve problems.

For additional information, see “Exception handling” on page 77.

TapeAlert

The TapeAlert facility in HP Ultrium drives allows applications to help avoid trouble by prompting the user to take remedial action, or in some cases, through the application automatically performing remedial actions itself.

For example, if the drive is experiencing trouble writing, the software can prompt the user to clean the heads, or, if there are several drives or an autoloader, automatically clean the heads without involving the user.

See “Monitoring the condition of the drive and media” on page 78 for more details.

Diagnostic logs

SCSI tape drives report problems in response to a REQUEST SENSE command from the host. If the backup application stores this information in a log file, it becomes significantly easier to troubleshoot problems, because the data can be used to pinpoint what is wrong.

Displaying drive information

Troubleshooting can also be simplified by giving users the ability to look at the drive’s firmware revision, and information about the host bus adapter. This information can be found by executing an

INQUIRY command, and can then be displayed, or stored in a log file.

Drive tests

A basic read/write test should be included in a backup application to check the integrity of the hardware. This should also allow the user to scan the SCSI bus and to solve problems concerning the device setup and configuration.

Design goals for LTO backup applications

• Use large SCSI read/write transfer sizes (256 KB is recommended).

• Incorporate data compression control and report the compression ratios achieved.

• Consider where to store directory information depending on the nature of the application.

• Only use immediate

mode.

• Use Cartridge Memory information to measure tape quality before backing up starts.

• Use the TapeAlert log to prompt the user to take remedial action to avoid problems.

• Use “cleaning required” indicators in the software to either prompt the user or enable the library

to use a cleaning cartridge to clean the drive heads.

• Allow users to set custom cleaning schedules.

WRITE FILEMARK commands, but avoid using other commands in immediate

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 13

HP restricted

• Use log files to store Inquiry and Sense Key/Error Code information about error conditions.

• Allow users to access drive firmware revision and HBA characteristic information

• Include the capability to download firmware.

• Incorporate simple diagnostic capabilities, such as Write/Read tests and SCSI device discovery.

• Incorporate online help.

Designing backup applications14

HP restricted

2 Configuration and initialization

This section covers the following topics:

• Operating System drivers

• Inquiry string recovery, finding information about the drive through the

• Additional LUN support, for operation with an autochanger device

• Fibre Channel support

Operating system drivers

Windows HP have a proprietary driver for Windows 2000 and Windows 2003. It is

intended that the driver is freely licensed to any software partner that requires it.

For the latest driver support for HP tape drives, please visit the following HP web site: http://www.hp.com/support/ultrium

NetWare HP has worked with Novell to provide driver support for HP’s Ultrium products. UNIX See the UNIX, Linux and OpenVMS Configuration Guide for details of how to

implement Ultrium support under the popular UNIX flavors.

Inquiry string recovery

HP Ultrium devices should not be recognized solely by the contents of their SCSI INQUIRY strings. In the past, hard-coded recognition of Inquiry strings has meant that software support for follow-on products from HP has been delayed when, to all intents and purposes, the new product was practically identical to the previous generation. For Ultrium, it is recommended that software applications ‘key off’ only the first eight bytes of the Product ID field—the text “Ultrium ”. Th e only use for the remainder of the bytes in this field is that they will be visible on-screen during the boot process of PC systems. As with HP’s DDS products, there will be very little difference between the first Ultrium drives and succeeding generations in terms of their basic SCSI characteristics; they will just store more data faster.

INQUIRY command

Standard INQUIRY Page Data SCSI SAS FC

Vendor ID (bytes 8–15) “HP ” “HP ” “HP ” Product ID (bytes 16–23) “Ultrium ” “Ultrium ” “Ultrium ” Product ID (bytes 24–31) “4-SCSI ” “4-SCSI ” “4-SCSI ” Product Revision Level (bytes 32–35) CRMV CRMV CRMV

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 15

HP restricted

Product ID, first 8 bytes

“Ultrium ” This will be the same for all HP Ultrium products, regardless of generation or model.

Product ID, last 8 bytes

1st byte: Generation identifier:

“4” LTO 4 (1600 GB at 2:1 compression) 2nd byte “–” Hyphen separator (ASCII 2Dh) 3rd–6th bytes “SCSI” SCSI protocol, regardless of transport or interface type

Product Revision Level

1st byte Product codename ID:

“A” “B”

“H”

LTO 4 SAS full-height drive LTO 4 SCSI full-height drive LTO 4 FC full-height drive

2nd byte Release type:

“0”, “1”

“2”

“3”, ...

Development Formal release

Post-release 3rd byte Minor release level: “0”–“9”, then “A” –”Z” 4th byte Firmware variant:

“D”

“W”

Standard distribution firmware

Standard HP automation firmware

Example

If new drive families or variants support features that are not available in previous generation products, you can detect the existence of these features through the SCSI

SENSE

commands. Exact details will become available as new products are defined. There is no

need to limit driver or application connectivity to a single HP Ultrium product type.

To determine the drive technology family:

Examine only the first eight bytes of the Product ID field (the text “Ultrium ”).

To determine the Ultrium format generation:

Use one of the following two methods, of which the second is preferred:

• Examine the character in byte 9. A “4” indicates format LTO 4 (1600 GB capacity at 2:1

compression) and so on.

• Preferred method: Use the SCSI REPORT DENSITY SUPPORT command.

For an LTO 4 product with LTO Ultrium 4 media, the following will be returned:

Primary Density Code: 46h Assigning Organization: LTO-CVE Linear Tape Open Compliance and Verification Entity Density Name: U-416 16 track

Configuration and initialization16

HP restricted

MODE SENSE and LOG

Since LTO 4 drives cannot write to Ultrium-2 media, the WRTOK bit will be clear for this media type.

Support for additional LUN

Enabling additional LUN support

When enabled by an internally-connected autochanger device, an extra Logical Unit Number (LUN) will be available at the target’s SCSI ID. This allows the attached autochanger device to be addressed via the tape drive. See ”Automation interface” on page 35. For ADI Bridging usage, the automation LUN will usually be LUN1.

No other LUNs are available on the drive, although HP is looking to provide new functionality through the use of additional LUNs in future products.

Supporting additional LUNs

When working with a library vendor who is incorporating HP Ultrium drives in products, software developers should liaise directly with the vendor about the functionality of the hardware available through the ADI or ACI.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 17

HP restricted

Configuration and initialization18

HP restricted

3Use of tapes

HP Ultrium user documentation and “Cartridges”, Chapter 9 of the Hardware Integration Guide, Volume 1 of the HP LTO Ultrium Technical Reference Manual, also contain information on cartridges.

Timing considerations are discussed in “Time-out values” on page 28.

LTO cartridge memory

NOTE: “Cartridge Memory” is the Ultrium version of the more general term “Media Auxiliary

Memory” or MAM, covering all media types.

Cartridge Memory has been added to the LTO cartridge for the following reasons:

• It speeds up load and unload times by removing the need to read system areas.

• It speeds up movement around tape by storing the tape directory (physical to logical mapping).

• It increases tape reliability because fewer tape passes are needed.

• It stores diagnostic and log information for tracking purposes.

Most of these uses are invisible to applications and handled internally by the drive. There is potential for applications to use the “Application Specific Data” area. This is being investigated.

For more details, see “Using Cartridge Memory” in “Using Special Features in Libraries”, Chapter 2 of the Hardware Integration Guide, Volume 1 of the HP LTO Ultrium Technical Reference Manual.

Identifying tape cartridge types

Using Cartridge Memory attributes

To identify the type of cartridge in the drive, read the Medium Type attribute in Cartridge Memory:

Attribute ID 0408h 00h

Using MODE SENSE

Examine the Medium Type field in the Mode Parameter header of the MODE SENSE command (byte 1 in the 6-byte version, byte 2 in the 10-byte version):

Medium Type 00h

01h 80h

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 19

Read/write (normal) data cartridge

01h

Cleaning cartridge

80h

WORM cartridge

Read/write (normal) data cartridge WORM cartridge SCSI and SAS only: CD (the drive is in CD-ROM mode)

HP restricted

Tape status and capacity

Following autoload or a LOAD command, the software can determine the state of the tape and its capacity from the Cartridge Memory and the Tape Capacity Log pages retrieved through the

SENSE

command. The information can also be invoked as a console operation at any time to find

the status and condition of the media.

Tape capacity figures can be used for two purposes:

• To give an application or user an indication of whether the tape has enough capacity for a

proposed backup. When using data compression, however, this is of little value, since the compression factor cannot be predicted accurately.

• Periodically during a backup to give an approximation of the amount of tape left.

CAUTION: An application should not use the capacity reported in the Tape Capacity log to fix the

backup size. This will result in permanent capacity truncation that could represent a significant percentage of the available capacity.

Finding the remaining capacity

Examine the Tape Capacity Log to estimate the effective remaining capacity of the tape (data-compression factors are not considered).

Interpreting Log Sense data

The following points affect the values returned in the data:

LOG

Units Capacities are given in megabytes (1,048,576 bytes) of user data and assume no

compression.

General • If data compression is used, the capacities are specified as though the drive is in

pass-through mode. The data compression factor is not considered.

• Regions of tape used by the system, such as EOD areas, are not included in

capacities specified. In other words, values are conservative.

• An allowance for read-after-write retries is made.

Maximum Capacity

Remaining Capacity

Use of tapes20

Maximum capacity values are only valid when the tape has completed a load sequence. If an immediate mode information until the tape has been successfully loaded and tape motion has ceased.

The remaining capacity value is the amount of tape remaining calculated from EOD. Remaining capacity values are only valid after the successful completion of the

following commands in non-immediate mode:

LOAD LOCATE MODE SELECT READ REWIND SPACE VERIFY WRITE WRITE FILEMARKS

The values after any subsequent command cannot be relied on unless the command is a sense type that does not cause any tape motion.

LOAD is made, LOAD SENSE will not return valid

HP restricted

Capacity calculations are based on estimates; reported values can be subject to error in two ways:

• Random errors caused by tolerances in tape length, hub diameter, and so on.

• Systematic errors caused by ignoring system areas, and so on. They ensure the calculated

capacity is actually available to the user. It is usually possible to write considerably more data than the calculated capacity.

Using the SET CAPACITY command

You can modify the capacity of a tape by changing the logical length of the tape through the SET CAPACITY command. The primary use envisaged is for testing purposes, although it may also be used in other circumstances where a shortened tape may be beneficial.

NOTE: All data currently on the tape will be lost following successful execution of this command.

The command is only accepted when the media is positioned at Beginning of Media (BOM).

With WORM cartridges, the command is only accepted and executed if the cartridge has not been initialized, that is, it has never been written to. Otherwise the cartridge is rejected with

CONDITION

attempted). TapeAlert flags 3Ch (WORM media—overwrite attempted) and 09h (write-protect) are set.

Command descriptor block

, sense key of Data Protect and additional sense of 300Ch (WORM media—overwrite

CHECK

76543210 0 Operation Code (0Bh) 1 Reserved (0) Immed 2 Reserved (0) 3 (MSB) 4 (LSB)

Capacity Proportion Value

5 Control

CDB fields

Immed 0 Status will not be returned until the

1 Status will be returned as soon as the CDB has been parsed.

Capacity Proportion Value

The portion of the total volume capacity to be made available for use. The value is the numerator of a fraction with a denominator of 65,535. The resulting

total volume capacity capacity proportion value×

available capacity will be .

------------------------------------------------------------------------------------------------------------------------------- -

Note that the LTO-4 format enforces a minimum tape length. A value that would result in a tape length below this minimum will be silently rounded up to the minimum permitted length.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 21

HP restricted

SET CAPACITY operation has completed.

65535

The following table gives the minimum acceptable Capacity Proportion Values and the approximate capacity they will give:

Cartridge Min. capacity proportion value Resultant approx. min. capacity Max. capacity

Ultrium 2 Ultrium 3 Ultrium 4

NOTE: Capacities are approximate and can be affected by defects that reduce the actual capacity

of the tape. Other factors, such as compression and block packing, may also affect capacity.

WORM media

HP Ultrium-3 WORM (Write Once—Read Many) data cartridges are two-tone grey/yellow to distinguish them from RW media. They have a unique Cartridge Type stored in the Cartridge Memory, so that they will be rejected by non-WORM compatible drives. For automation configurations with auto-eject disabled, the cartridge will not be physically ejected from the drive but held at the “ready eject” position.

The write-protection tab behaves as on Read/Write (RW) cartridges.

How WORM media works

Drives use the EOPD (End of Protected Data) value to control the use of WORM tapes.

1605h 17.2 GB 200 GB

151Ah 33 GB 400 GB

1055h 51 GB 800 GB

EOPD is a logical position on tape that is automatically calculated based on the End of Data (EOD) value read from the Cartridge Memory (CM) when the cartridge is loaded into the drive. The EOD value is an “intrinsic” code stored and protected in the Cartridge Memory of each WORM cartridge, and updated after each write session. The EOPD indicates that data between BOM and this position cannot be overwritten.

The EOPD value is held within the drive’s memory. It is updated automatically and continuously as each block of data (typically 64 or 128 KB) is written to tape, so the EOPD value indicates a logical position immediately after the last block of data written to tape.

When the cartridge is unloaded, the drive updates the EOD value in CM to reflect the end of successfully written data on the cartridge, and clears the EOPD value stored within the drive. Any future writes to the cartridge will occur after the location of the EOD, which will become the initial location for EOPD during the next write operation.

Changes to SCSI commands

New additional sense codes and TapeAlert flags

ASC/Qs:

• 300Ch (WORM medium—overwrite attempted)

• 300Dh (WORM Medium—integrity check failed)

Use of tapes22

HP restricted

TapeAlert flags:

• 3Bh (WORM medium—integrity check failed)

• 3Ch (WORM medium—overwrite attempted)

Error Usage page

For WORM cartridges, the Wrap Number fields in the Error Usage page are replaced by an Error Code field. This contains the ASC/Q value reported to the host when the associated error was detected.

Only appended writes accepted

If a WORM cartridge is placed in a WORM-compatible drive, the drive will accept write commands (records, filemarks) only if the current logical position is beyond the position identified by the EOPD value. If a write command is received by the drive when the logical position is before the EOPD value, the command will be rejected and Protect and additional sense of 300Ch (WORM medium—overwrite attempted). The TapeAlert flags 3Ch (WORM medium—overwrite attempted) and 09h (write-protect) are set.

CHECK CONDITION returned with sense key of Data

ERASE commands rejected

ERASE commands (short or long) to a drive containing a WORM cartridge will not overwrite or erase user data on tape. Any tape is rejected and sense of 300Ch (WORM medium - overwrite attempted). The TapeAlert flags 3Ch (WORM medium—overwrite attempted) and 09h (write-protect) are set.

CHECK CONDITION returned with sense key of Data Protect and additional

ERASE command that would result in user data being over-written on

SET CAPACITY command

The SET CAPACITY command will only be accepted and executed if the WORM cartridge has not been initialized, that is, it has never been written to.

If a

SET CAPACITY command is received by the drive when the cartridge has been initialized, it is

rejected and 300Ch (WORM medium—overwrite attempted). The TapeAlert flags 3Ch (WORM medium—overwrite attempted) and 09h (write-protect) are set.

CHECK CONDITION returned with sense key of Data Protect and additional sense of

Re-writing media labels

If there is no user data on the tape, the media label can be rewritten. The label contains software application-related information such as a unique identification code and does not contain user data.

Writing is allowed when the current logical position is at BOT and:

• there are only filemarks between this position and EOD, or

• there are only 1 or 2 sequential records followed by any number of filemarks, but no further

records, between this position and EOD.

Allow overwrite of last filemarks before the EOD data set

At the end of a backup or archive session, many software applications write two filemarks to tape immediately before the EOD data set is written. These filemarks are logical markers that enable the

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 23

HP restricted

application to locate the tape in preparation for subsequent writing or reading operations. At the start of an appending archive or backup session, it is common for the application to locate the tape to a logical position immediately preceding the second filemark and to overwrite the second filemark during the data appending session.

Writes are allowed:

• when the current logical position is at EOD. This means that the drive must have read the EOD

from tape before attempting to overwrite it.

• when there are only filemarks between the current logical position and EOD, and at least one

filemark immediately before the current logical position.

Using CM to check tape integrity

Drives that support WORM cartridges check that the anti-tampering measures have not been violated before and during media access operations. If a violation is detected, for example, if the content of the CM does not match the content of the tape, the cartridge is treated as read-only, and Tape Alert flag 3Bh (WORM medium—integrity check failed) is set.

HP strongly recommends that software applications check for the presence of TapeAlert flag 3Bh after a tape load and periodically during operation. If the flag is set, the software should alert the operator and log the incident for audit.

Hosts can use the WTRE bit on the Device Configuration mode page (bit 6, byte 15 of mode page 10h) to control the behavior of the drive when reading WORM media whose WORM integrity is in doubt and which may have been tampered with. See details of the mode page in Chapter 4 in SCSI Interface, volume 4 of the HP LTO Ultrium Technical Reference Manual for more information.

Behavior with a missing or inconsistent EOD value

A missing EOD value or one in which the value on tape differs from that in CM can be caused by:

• CM corruption

• deliberate, malicious alteration of the EOD value in the CM

• an interruption of the drive power supply while writing data.

If the CM indicates that the EOD is not valid (for example, if the drive has powered down during a write), TapeAlert flag 04h (Media) will be set on cartridge load.

The drive behaves as for a RW cartridge with no EOD data set, except that any attempt to overwrite data will be rejected with C 2700h (write-protected). Tape Alert flag 09h (Write-Protect) is set.

HECK CONDITION, a sense key of Data Protect and additional sense of

Unique media identifier

For added security, HP strongly recommends that applications read and track cartridge manufacturer and serial number values from the Cartridge Memory using MAM access commands:

• Attribute ID 0x0400: Cartridge manufacturer ID

• Attribute ID 0x0401: Cartridge serial number

Both values should be read and concatenated to ensure that the number is unique.

Use of tapes24

HP restricted

Barcode support

Ultrium barcode support is required for WORM media so that the application and tape library can distinguish WORM media from normal RW media or cleaning cartridges.

HP recommends the use of the following barcode formats for Ultrium media:

123456L2 LTO Ultrium 2 123456L3 LTO Ultrium 3 123456L4 LTO Ultrium 4

123456LT LTO Ultrium 3 – WORM

123456LU LTO Ultrium 4 – WORM

Responding to Cartridge Memory data

NOTE: Software should use the TapeAlert log in preference to the Cartridge Memory to detect

conditions which require the user or host to take preventative action. See “Monitoring the condition

of the drive and media” on page 78.

These guidelines indicate how host applications should make use of the data contained in the Cartridge Memory during normal operation (that is, when tapes are not permanently write-protected, not constantly re-formatted).

The console messages triggered by these criteria should clearly indicate a course of action to the end-user, such as the following:

1. Clean the tape heads using a cleaning cartridge.

2. Insert a new tape cartridge.

3. Archive the data.

Load count

NOTE: This only applies when non-write-protected cartridges are used.

The load count is the number of times the cartridge has been loaded into a drive and accessed.

Hewlett-Packard recommends a maximum use for a tape of 20,000 passes over any particular area of the tape. This conservative estimate is also influenced by the quality of the application and the driver software in being able to maintain streaming, thereby preventing repositioning over the same area of tape, without data being transferred.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 25

HP restricted

RWW retry counts

Data is read immediately after being written to tape to establish that it has been written correctly. Increases in RWW retries can be due to four factors:

• Deterioration in the media

• Dirty heads

• Drive malfunction

• The operating environment

Corrective action

The recommended criteria for corrective action are as follows:

RWW Retries > 5% Total data sets written

When using tapes without write-protection, use the Total count.

The corrective action should be as follows:

1. Use another tape and, for a write operation, try repeating the write. For a read operation, try

reading data from the tape.

2. See whether the current RWW value is within the recommended limit.

3. If the values are now within the limit, you can assume that the original tape is nearing the end of

its useful life. Proceed as follows:

• For a write operation, discard the tape and use a new one.

• For a read operation, transfer the data to a new tape.

4. If the value is still outside the limit, clean the tape heads with a cleaning cartridge and try

repeating the operation with the original tape.

Use of tapes26

HP restricted

4 Factors affecting performance

This chapter contains techniques and information to help you design software applications so that they use the tape drive’s potential as efficiently as possible.

• Ways of optimizing performance:

• Ensuring the recommended minimum transfer sizes page 27

• Identifying the media type page 28

• Using Cartridge Memory instead of tape headers page 28

• Using the Performance Log page to diagnose problems page 28

• Time-out values to help you tune timings in backup applications page 28

• Log pages—recommended support page 29

• Factors affecting performance, relating separately to the host, drive and format page 29

Ways of optimizing performance

HP’s Ultrium drives are high-performance products. Application software may require significant enhancement in order to capitalize on this speed. There are a number of areas to look at and these are discussed below.

Further details can also be found in the “How to optimize the performance of hp ultrium tape drives” white paper.

Detecting the drive’s speed

Applications should not key off Inquiry strings in order to tell the difference between different speed drives. It is better to use the Performance Log page see under the 4, “Commands”, of SCSI Interface, Volume 3 of the HP LTO Ultrium Technical Reference Manual.

In the Performance Log page (34h), parameter 04h (Native data rate) gives the native data rate of the drive in units of 100 KB/s. LTO 4 drives give the value 04B0h, indicating 120 MB/s with Ultrium 4 media or no cartridge loaded. If a previous generation cartridge is loaded, the value will be lower.

Ensuring the recommended minimum transfer sizes

Use the Data Compression Log page. HP cannot diagnose performance issues without accurate reporting of the current compression or the average compressibility over a backup session. Make sure that you report the log page.

Regarding HP’s One-Button Disaster Recovery (OBDR) feature (see “One-Button Disaster Recovery

(OBDR)” on page 89), it is important to note that in some situations the SCSI block size may have to

be fixed for a given tape for format reasons. This means that if the host writes 2 KB blocks to support OBDR, it may have to continue to write 2 KB blocks for the rest of the tape; it depends on the format compatibility required by the overall system. However as HP Ultrium drives are insensitive to

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 27

HP restricted

LOG SENSE command in Chapter

absolute block size, performance should not suffer, but do ensure that the transfer size is at least

256 KB.

Maximum block size

The READ BLOCK LIMITS command indicates that block sizes and variable length transfer sizes are supported for values between 1 byte and 16,777,215 bytes.

Media type identification

HP recommends that you use the REPORT DENSITY SUPPORT command (with the Media bit enabled) to identify the type of media loaded in the drive.

Using Cartridge Memory instead of tape headers

For optimum performance, it is also important that the host writes application tape header information to the Cartridge Memory (see “Cartridge Memory (LTO-CM)” on page 35) rather than to the actual tape. This allows cartridges to load and unload quickly and prevents excessive media wear at the beginning of the tape. As the access method to Cartridge Memory data is an open standard, it also permits other software systems to identify alien media positively in shared storage environments.

Using the Performance Log page for diagnosing problems

The Performance Log page (34h) contains data that should allow application software to monitor the data-rate being sent to the drive dynamically. For details, see the Volume 3, SCSI Interface, of the HP LTO Ultrium Technical Reference Manual.

LOG SENSE command in

Time-out values

SCSI Command Recommended Time-Out Value

Load 10 minutes Unload 10 minutes Rewind (full tape length) 10 minutes Space/Locate/CD-ROM Read (10) 20 minutes Erase (long) 5 hours Erase (short) 5 minutes Write/Write Filemarks 5 minutes Read 20 minutes Read/Write Attribute (MAM), with 1 KB of attribute data 1 minute Non-tape movement (such as TEST UNIT READY, INQUIRY) 1 minute

Notes:

• These values are for a single SCSI command in non-Immediate mode. As most commands will

be sent in Immediate mode, status will be received by the host typically within 20 ms. In such cases, the time-out given indicates when the drive will have completed the operation and be ready for the next tape movement command.

Factors affecting performance28

HP restricted

• All of these values may be subject to change.

• There is no retension facility.

Recommended support of log pages

Some of the media-related data items on the log pages are duplicates of data that is available through the specification. We recommend that you use MAM commands as the primary source for such data, because this access method is portable to tape drives from other vendors, that is, the data is not in a vendor-unique format.

In the long term, HP intends to expose all tape usage and drive hardware usage information via the industry-standard MAM-format commands, so it is wise to start to implement this approach now.

READ ATTRIBUTES command using the Media Auxiliary Memory (MAM) access

For full details of the Ultrium log pages, see the of SCSI Interface, Volume 3 of the HP LTO Ultrium Technical Reference Manual.

Factors affecting performance

Further details on improving performance can be found in the “How to optimize the performance of hp ultrium tape drives” white paper.

Host-related factors

Performance Factor Detail

Host SCSI performance The execution of each SCSI command involves a number of bus

phases, of which the data phase is only one. The key phases are as follows:

• Intra-command bus-free time

• Arbitration and selection

• Message out

• Command

Host Burst Rate During the data phase of each SCSI command, data is transferred to

or from the drive at the host’s burst rate. If the host’s burst rate is slow, then it takes longer to transfer the data. Extra time during this phase is simply added to the total command time, and so it can affect the overall performance.

Even if the burst rate is much faster than that required to maintain streaming, the total command time may prevent the commands from being issued fast enough.

LOG SENSE command in Chapter 4, “Commands”,

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 29

HP restricted

Performance Factor Detail

Example: SCSI: Consider an 8 KB transfer at burst rates of 8 MB/s and 1 MB/s.

The fast transfer takes 1 ms, while the slow transfer takes 8 ms. Since the rest of the command may only take 4–5 ms, the difference of 7 ms is very significant.

SAS and FC: The SAS link is capable of 3 Gb/s (300 MB/s) whereas

the performance of the FC drive depends on the negotiated link speed of the host/HBA port, any intermediate switch ports and the drive port. The ports will negotiate to the highest possible mutually supported speed, ideally 4 Gb/s (400MB/s), then 2 Gb/s (200MB/s) or as low as 1 Gb/s (100MB/s).

Disk Subsystem Performance

The speed and configuration of the disks used will have a significant impact on the backup speed of the whole system.

Recommendation: Using RAID can have a significant effect on the throughput of the

whole system, by the use of interleaved disk reads. Use more spindles where possible or a reasonably sized RAID system. More disks means more throughput.

Note that the Raid level makes a difference to performance. RAID5 will be slower for writes (restores) than reads (backups). RAID0 and RAID1 are faster but expensive in terms of numbers of spindles required and not so tolerant to disk failure (RAID0 is particularly intolerant).

File System Efficiency Operating systems vary in the efficiency with which they retrieve files

sequentially for backup applications. Most operating system development effort is put into speeding up access times within files rather than file seek times.

Consider using staging technology to stage an image before writing to tape, particularly with many small files, which will impact performance due to file accession.

Hardware Configuration If the disk and tape drives are on separate buses, the effective

available bandwidth can be doubled.

Recommendation: Use one HBA for disks, and put the tape drive on a separate bus. It

makes sense to split heavily used FC cards across separate PCI busses so that they do not contend for PCI bus bandwidth.The more PCI busses the better.

Host CPU Speed Faster hosts can typically transfer data quicker.

Recommendation: Use as fast a processor as possible for the backup system.

Network Transfer Time If backup involves transferring data over the network, network

performance is often a major bottleneck.

Factors affecting performance30

HP restricted

Performance Factor Detail

Recommendation: • Try not to put Ethernet in the way of data transfer unless staging

Write Commands Do not interleave write commands with other commands, such as

Drive-related factors

Performance Factor Detail Drive’s SCSI

Performance

Recommendation: The host must always ensure the following:

Transfer Mode The transfer mode can be fixed or variable, selectable through the

technology is being used; aggregation of multiple Ethernet clients remains a good strategy to delivering on drive performance.

• Even with Gbit Ethernet, the effective throughput is less than that of

an LTO Ultrium 4 drive, so either use carefully designed topologies, or stage an image first and use locally attached tape, otherwise the Ethernet itself can become the bottleneck.

Gbit Ethernet can be used in a carefully designed topology. Alternatively, stage an image first and then use locally attached tape for the backup.

READ

POSITION

TapeAlert log page during a long write.

In order to minimize SCSI bus loading, the drive must execute its SCSI phases quickly. The phases are as follows:

• Selection

• Message-out identification

• Receipt of the command

• Disconnection

• Mid-command bus-free time

• Arbitration and reselection

• Message-in identification

• Disconnects are enabled

• Synchronous negotiation is enabled and established between the

• The drive is in buffered mode

• When reading and verifying, always use the same block size as

MODE SELECT command. Fixed Mode: The transfer size is equal to the (block) size multiplied by

the number of records (blocks) in the transfer.

and LOG SENSE. Do not, for example, attempt to read the

drive and the HBA

that in which the tape is written, otherwise performance will be very seriously affected.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 31

HP restricted

Performance Factor Detail

Recommendation: A good transfer size to aim at is 256 KB (128 KB minimum). For an

application that uses 512-byte records, each fixed-mode transfer should transfer 512 records. Higher transfer sizes are also recommended for higher compression ratios.

Variable Mode: Only one block is transferred at a time. The size of the block determines the size of the transfer. Ideally the application should

aim to use 256 KB blocks.

Records (Block) Size The size of the transfer impacts the performance, rather than the size of

the record (blocks) in the transfer.

Recommendation: As above, aim to use 256 KB transfers. Higher transfer sizes are also

recommended for higher compression ratios.

Transfer Size Transfer size is the amount of data transferred for a single command,

whether the drive is in fixed or variable block mode. In both fixed and variable modes, the drive works best if it receives a

large amount of data for each command, so a large transfer size for write commands is recommended.

At small block size, the transfer rate is substantially degraded. This is because the drive controller and the host spend too much time handling SCSI overhead instead of writing data to tape, resulting in stream-fails. The block size at which this happens varies between drives, but generally the faster the drive, the larger the block size needed to stream.

Recommendation: Use 256 KB transfers as a minimum. Higher transfer sizes are also

recommended for higher compression ratios.

Transfer Direction There are some noticeable performance differences between reads

and writes, caused by the extra device CPU time needed by the drive to read data sets from the media.

Recommendation: Use large transfer sizes; the drive is less likely to stream-read small

transfer sizes than it will when writing transfers of the same sizes.

Format-related factors

Performance Factor Detail Tapemarks Tape marks (filemarks) have many different uses to give a logical

Recommendation: Write filemarks as rarely as is reasonable for your application.

Factors affecting performance32

structure to data on a tape. The SCSI Standard specifies certain actions that the drive must take when it is told to write a filemark.

If the drive is told to write a filemark when the Immediate bit is not set, the standard insists that the drive must flush all data to tape. If used unnecessarily this will adversely affect performance and waste tape capacity.

HP restricted

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 33

HP restricted

Factors affecting performance34

HP restricted

5 Supporting Ultrium features

This section covers the following features of HP Ultrium drives:

• LTO Cartridge Memory page 35

• Automation and drive interface (ACI and ADI) page 35

• Cleaning page 50

• Resetting drives page 50

• Backup software page 50

• Controlling data compression page 51

• Other Mode page information page 51

Cartridge Memory (LTO-CM)

LTO Cartridge Memory (LTO-CM) is EEPROM memory that is embedded in every LTO Ultrium tape cartridge. It is non-volatile and is contactless in that it is read by RF coupling rather than electrical contact.

Further information

• For general information about LTO-CM, see “LTO Cartridge Memory” in Chapter 5, “Cartridges”

in Background to Ultrium Drives, Volume 6 of the Ultrium Technical Reference Manual.

• For suggestions of how to make use of cartridge memory in libraries, see “Using Cartridge

Memory” in “Using Special Features in Libraries”, Chapter 2 of the Hardware Integration Guide, Volume 1 of the HP LTO Ultrium Technical Reference Manual.

Automation interface

The interface between the tape drive and the library will support both the open standard Automation/Device Interface (ADI) and the proprietary Automation Control Interface (ACI). The tape drive defaults to the ACI protocol and will only transition to ADI when the library initiates an ADT login exchange.

NOTE: The same connector is used for both interfaces; for convenience, it is referred to as the ADI

Connector in this manual.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 35

HP restricted

Automation/Device Interface (ADI)

There are two elements to the Automation/Device Interface (ADI):

• ADI Transport Protocol (ADT)—a standard protocol for communication between a SCSI

automation device and a SCSI data transfer device, such as a tape drive. The ADT protocol allows conforming ADI SCSI devices to inter-operate. The objectives of ADT

are:

• To provide a low-cost interconnect method between an automation device and the data transfer devices that reside within the media changer.

• To standardize this interface so that different disk drives, tape drives, optical media drives, and other SCSI devices may be added to conforming media changers without requiring modifications to generic system hardware.

• To provide for the addition of special features and functions through the use of vendor-specific options.

The interface protocol includes provision for the connection of two SCSI ports. One of these ports is intended to be attached to a media changer device and may operate either as a SCSI initiator port or a SCSI initiator/target port. The other port is intended to be attached to a data transport type device (tape drive) and may operate as either a SCSI target port or SCSI initiator/target port. No provision is made for connection of more than two ports.

• ADI Commands (ADC-2)—an extension to the SCSI command set for communication with

application clients through the ADI. The objectives of ADC-2 are:

• To permit an application client to communicate over a SCSI service delivery subsystem, with a logical unit that declares itself to be an ADI device in the Peripheral Device Type field of the standard

• To define commands unique to the ADI device type.

• To define commands and parameters to manage the operation of the ADI device type and the operation of logical units of other specific device types that are present in the same device as the ADI logical unit.

INQUIRY data.

For details of HP’s implementation of these standards, see the HP ADI Firmware Integration Guide. The T10 standards referenced by this guide are as follows:

• ADT: ANSI INCITS 406-2005

• ADC-2: T10/1741-D revision 7d.

Supporting Ultrium features36

HP restricted

Automation Control Interface (ACI)

The Automation Control Interface (ACI) protocol allows the activities of the drive to be coordinated within a library. The protocol has been designed so that it can be made into a standard feature of tape drives. It provides a rich and extensible functionality to allow automation manufactures to add value in their application of it.

The interface is a serial bus with additional control lines, designed to connect the Ultrium tape drive to an automation controller in a tape library. Each tape drive position has a separate ACI connection allowing communication to the drive via its RS 422 serial port.

The ACI protocol provides the following fundamental functions:

• Coordinating the automation controller and the tape drive for Load and Unload operations

• Allowing the automation controller to retrieve information from the tape drive

• Setting tape drive configuration information

In addition, the following functions may be supported depending on the way that the tape library is configured:

• Providing upload and download of firmware images

• Providing access to Cartridge Memory contents

• Providing a protocol for passing SCSI commands to the tape drive over the interface.

The ACI protocol allows for “packetized” SCSI commands to be sent from the attached controller and submitted to the tape drive as if they have been received on the drive’s own SCSI bus. For example, the ACI protocol allows the sending of load/unload commands from a specially defined automation command set to cause drive action. This ability enables the attached controller to access and control the drive in exactly the same way as it would over the SCSI bus.

NOTE: Ultrium SCSI Parallel drives implement a limited subset of the SCSI commands, including

INQUIRY, LOG SENSE, LOG SELECT and MODE SENSE.

The following notes should be regarded as supplementing the ACI protocol specification rather than replacing them. Please refer to the specification for further implementation details. These notes refer to the “standard” automation drive variant. Different behavior may be exhibited in certain areas for specific OEM variants as requested by OEMs.

Modes of usage through ACI

Slave to a library controller

The ACI can receive commands such as LOAD and UNLOAD from a specially defined automation command set to control the action of the drive.

Most tape libraries need to have a means of communication between controller and tape drives, to enable correct synchronization of mechanical operations between drive and picker arm. For instance, in a soft load capable device such as an Ultrium drive, the picker must let go of the

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 37

HP restricted

cartridge at the moment that the drive starts to pull it into the drive, if it is configured for that type of operation.

This degree of control over synchronization cannot be achieved though the host’s backup software; it must be controlled directly by the library controller. Most tape libraries work this way today. The process is transparent to the backup software.

SCSI pass-through mode

The ACI protocol allows the transfer of “packetized” SCSI commands from an attached controller and submit them to the tape drive as if they have been received on the drive’s own SCSI bus. This enables the attached controller to access and control the drive in exactly the same way as it would via the SCSI bus.

ACI command set

The following ACI commands are supported on HP Ultrium drives:

Mandatory Commands Optional Commands 00h Get Drive Info 40h Send SCSI Command 01h Load 42h Send Firmware Image 02h Unload 43h Get Firmware Segment 03h Get Drive Status 49h Get Buffer Size 04h Set Drive Configuration 4Ah Send Firmware Segment 05h Get Drive Configuration 4Bh Set Time 06h Reset 4Ch Get Time 07h Set Baud Rate 08h No Op 09h Get Error Info

0Ah Acknowledge Attention

ACI commands that affect drive streaming performance

Commands that alter the state of the drive in some way will affect the performance of the drive when stream reading or writing. It is recommended that no command within the following set are sent to the drive while the drive is writing or reading as it would affect the data throughput to or from the drive:

• Load

• Unload

• Send Firmware Image

• Send Firmware Segment

• Reset

• Set Drive Configuration—if the host (SCSI or FC) is reconfigured

Supporting Ultrium features38

HP restricted

• Send SCSI with the following opcodes:

Log Select Mode Sense Read Attribute Mode Select Request Sense Write Attribute

New features in ACI 4.3

The following sections describe the differences between revision 4.2 of the ACI specification, used in HP LTO Ultrium 3 drives, and revision 4.3, used in HP LTO Ultrium 4 tape drives.

The main additions to ACI 4.2 revision are:

• Improved backward compatibility

• Support for encryption

Backward compatibility

To provide backward compatibility with ACI 4.0, SCSI parallel drives initialize in ACI 4.0 compatibility mode. Similarly, Fibre Channel drives initialize in ACI 4.1 compatibility mode and SAS drives initialize to ACI 4.2 compatibility mode. This means drives with ACI 4.0, ACI 4.1 or ACI

4.2 can be replaced with drives with ACI 4.3, without the need to replace or update the automation controller firmware.

• In ACI 4.0 compatibility mode the drive will accept Set Drive Configuration CMD_DATA in ACI

4.0 format and will respond to Get Drive Status and Get Drive Configuration with ACI 4.0 RDATA.

• In ACI 4.1 compatibility mode the tape drive will accept Set Drive Configuration in ACI 4.1

format and will respond to Get Drive Status and Get Drive Configuration with ACI 4.1 RDATA.

• In ACI 4.2 compatibility mode the tape drive will accept Set Drive Configuration in ACI 4.2

format and will respond to Get Drive Status and Get Drive Configuration with ACI 4.2 RDATA.

An ACI compatibility mode does not restrict the use of ACI 4.3 commands or the use of new fields to earlier ACI version commands and will respond with the appropriate ACI 4.3 RDATA in these cases.

• A parallel SCSI tape drive will remain in ACI 4.0 compatibility mode until it receives a valid ACI

4.1 format Set Drive Configuration command or an ACI 4.2/4.3 format Set Drive Configuration command with the ACI Major/Minor Version fields set to a new and valid ACI version (which is 42h). It will then respond with the appropriate ACI version RDATA. The drive will return to ACI

4.0 compatibility mode in the following circumstances:

• If the tape drive receives a valid ACI 4.0 format Set Drive Configuration command.

• An ACI 4.2 or ACI 4.3 format Set Drive Configuration command with the ACI Major/Minor Version fields set to 40h.

• The tape drive is reset.

• A Fibre Channel tape drive will remain in ACI 4.1 compatibility mode until it receives a valid

ACI 4.2/4.3 format Set Drive Configuration command with the ACI Major/MinorVersion fields set to a new and valid ACI version (which is 42h). It will then respond with the appropriate ACI version RDATA. The drive will return to ACI 4.1 compatibility mode in the following circumstances:

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 39

HP restricted

• If the drive receives a valid ACI 4.1 format Set Drive Configuration command.

• An ACI 4.2 or ACI 4.3 format Set Drive Configuration command with the ACI Major/Minor Version fields set to 41h.

• The tape drive is reset. Fibre Channel tape drives do not support ACI 4.0 compatibility mode, so they will only accept ACI version 4.1 and later commands.

• An SAS tape drive will remain in ACI 4.2 compatibility mode until it receives a valid ACI

4.2/4.3 format Set Drive Configuration command with the ACI Major/Minor Version fields set

to a new and valid ACI version (which is 43h). It will then respond with the appropriate ACI version RDATA. The drive will return to ACI 4.2 compatibility mode in the following circumstances:

• If it receives an ACI 4.2 or ACI 4.3 format Set Drive Configuration command with the ACI Major/Minor Version fields set to 42h.

• If the drive is reset. SAS tape drives do not support ACI 4.0 or ACI 4.1 compatibility mode, so they will only accept ACI version 4.2 and later commands.

HP LTO tape drives support all ACI versions later than their default ACI compatibility mode version. For example, Fibre Channel drives default to ACI 4.1 compatibility mode but also support ACI version 4.2 and ACI 4.3, which can be selected using the appropriate format ACI Set Drive Configuration command.

Notes

An HP LTO tape drive supports a two-step initialization sequence. The first step behaves as described above; the drive goes through the first step of the initialization sequence and then begins to send the <ENQ> character at 10-second intervals. The drive sends the <ENQ> character within 500 ms of the power-on, drive reset, ACI reset or completion of firmware upgrade.

Once the drive sends the first <ENQ> character, it will begin the second step of initialization. This may take several minutes to complete as it includes rewinding the tape and, optionally, unthreading it, if the drive contains a seated cartridge with threaded media.

During the second step of initialization, the drive will respond to all ACI commands, except Get Drive Info, Get Error Info and some Send SCSI commands with BUSY status. The drive will respond normally to the Get Error Info command. It will respond normally to the Get Drive Info command except that every byte of the Manufacturing Data Code and Serial Number fields will contain the value FFh. During initialization the drive supports SCSI commands as in the following table. It responds with BUSY status to commands that are unsupported:

SCSI command Supported during initialization

INQUIRY LOG SELECT LOG SENSE MODE SELECT MODE SENSE READ ATTRIBUTE RECEIVE DIAGNOSTIC REPORT

Supporting Ultrium features40

HP restricted

Yes No Yes No No No No

SCSI command Supported during initialization

REQUEST SENSE REPORT DENSITY REPORT REPORT LUNS SEND DIAGNOSTIC TEST UNIT READY WRITE ATTRIBUTE

Yes No Yes No Yes No

When the drive has completed the second step of initialization, it will respond normally to all supported commands, and it will report the correct Manufacturing Date Code and Serial Number in the corresponding fields of the Get Drive Info RDATA.

HP recommend that automation controllers use the Get Drive Status command to detect the completion of the second step of tape drive initialization. After the drive sends the first <ENQ> character and until it completes the second step of initialization, it will respond to a Get Drive Status command with BUSY status. After the drive has completed the second step of initialization, it will respond normally to the Get Drive Status command.

Encryption support

For parallel SCSI and SAS drives, the RDATA returned by the Get Drive Status command contains an Encryption Status field:

76 5 4 32 1 0

0 Vendor

Unique

1 Cartridge Type Compression Tape Activity 2 Drive

Error

3–4 Reserved

5 Reserved Encryption Status WORM

Cartridge

Present

Media

Error

Prevent Media

Removal

Write Protect

TapeAlert Rsvd

Ready

Eject

Clean

Expired

Ready

Access

Cleaning

Cartridge

Load

Clean

Required

Ready

Load

Clean

Needed

For FC drives, the RDATA returned is in the following format:

76 5 4 32 1 0

0 Vendor

Unique

Cartridge

Present

Prevent Media

Removal

Write Protect

Ready

Eject

Ready

Access

1 Cartridge Type Compression Tape Activity 2 Drive

Error

3 Port 0

Active

4 Port 1

Active

Media

Error

TapeAlert Reserved

Clean

Expired

Cleaning

Port 0 Topology Port 0 Speed

Port 1 Topology Port 1 Speed

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 41

HP restricted

Cartridge

Load

Clean

Required

Ready

Load

Clean

Needed

76 5 4 32 1 0

5 Reserved Encryption Status WORM

The Encryption Status bits indicate the current encryption/decryption status of the SSC device server:

Value Meaning

00b The device server is currently not performing encrypting or decrypting operations. 01b The device server is currently performing encrypting or decrypting operations.

10b, 11b Reserved.

Further details

For more information about ACI, see the ACI Specification supplied by HP.

Supporting the ACI protocol

Software vendors implementing support for attached library devices will need to work closely with the library vendor concerned. See HP’s “Ultrium Automation Cookbook” for more details.

Recommended ACI time-out values

ACI commands fall into three broad classes:

• Commands that the drive executes immediately

• Commands that the drive queues but which it can execute concurrently with auto-mode reads

and writes (in other words, streaming operation)

• Commands that the drive queues but which interrupt streaming operation.

The response time to an ACI command will depend on the type of ACI command and the activity status of the drive at the time the command is received.

Note that the drive does not support ACI command queuing. However, under exception conditions, command queuing may occur, say if the automation controller had timed-out the tape drive’s response to a command and either resent the command or sent another command. In these circumstances, the drive will not ignore the overlapped commands but will respond to every command package it had received.

An example of when this may occur if the host issued a long SCSI the automation controller issued an

UNLOAD command until the long erase had completed. If the automation controller timed out the

drive’s response to the

UNLOAD command and re-sends the command or sends another command,

UNLOAD command, then the drive would not respond to the

then it needs to be able to handle the response to the original

ERASE command to the drive and

UNLOAD command as well as to the

subsequent commands.

The following tables list the recommended ACI command time-outs for queued and non-queued commands.

Supporting Ultrium features42

HP restricted

Table 1 Non-queued ACI commands

ACI Command Recommended time-out value

Get Drive Info 5s Get Drive Status 5s Get Drive Configuration 5s Get Error Info 5s Get Buffer Size 5s No Op 5s Acknowledge Attention 5s

Table 2 Queued ACI commands

ACI Command Recommended time-out value

ACI Load—immediate 5s ACI Load—non-immediate (drive idle, unloaded) 300s ACI Unload—immediate 5s ACI Unload—non-immediate (tape loaded, at EOM,

drive idle) Set Drive Configuration (tape loaded, at EOM, SCSI

unload) ACI Reset—ACI bus 5s ACI Reset—drive 5s Set Baud Rate 5s

300s or 9000s depending on

implementation strategy

300s or 9000s depending on

implementation strategy

Treatment of reserved fields

To ensure forwards compatibility with future versions of the ACI, automation controller firmware should set any command fields labelled as ‘Reserved’ to zero. Likewise, automation controller firmware should mask off any response fields labelled as ‘Reserved’ during the processing of tape drive responses. This will allow older versions of automation controller firmware to operate successfully with newer versions of tape drive firmware.

Recommended power-up sequence

After power-up, HP recommends that the automation controller wait until it has received at least one ASCII <ENQ> character from the tape drive before attempting a command-response transaction. HP Ultrium tape drives use a two-step power-up sequence and the drive sends <ENQ> to signal the transition between the steps. The drive sends the first <ENQ> within 500 milliseconds of exiting the reset state after receiving power. Note that the default baud rate at power-up is 9600.

Consider sending Get Drive Info as the first command, either packetized or primitive. This retrieves a variety of useful identifying information identifying the tape drive, including the version of the ACI protocol that the tape drive supports.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 43

HP restricted

During the second step of the power-up sequence, the tape drive will respond with BUSY status to all ACI commands except Get Drive Info and Get Error Info. The amount of time taken by this second step will vary widely depending on three factors:

• The presence or absence of a cartridge in the tape drive

• The position of the media if a cartridge is present

• The ability of the tape drive to access the cartridge memory if a cartridge is present

HP recommends that the automation controller polls using the Get Drive Status command to monitor the completion of the power-up sequence. When the tape drive returns

GOOD status to a Get Drive

Status command, it has completed the power-up sequence.

If operating with a tape drive that supports ACI V4.1 and higher, HP recommends that the automation controller synchronizes the tape drive’s time stamping clock to its own using the Set Time command once the tape drive has completed the power-up sequence.

In some circumstances when responding to the first Get Drive Info command, the tape drive will fill every byte in the Manufacturing Date Code and Serial Number fields with FFh. The tape drive behaves this way when it receives the Get Drive Info command during the second step of the power-up sequence because it cannot access the EEPROM that stores this information at that time. The automation controller may retrieve the correct value for these fields with a second Get Drive Info command sent after the power-up sequence completes.

Once the power-up sequence completes, the automation controller can configure the tape drive using the Set Drive Configuration command. Each time a Set Drive Configuration command is sent, it is recommended that a Get Drive Configuration command is sent to double-check that the drive is configured correctly.

It is recommended that the Get Buffer Size command is sent to drive as part of the power-up sequence to determine the maximum burst buffer size and maximum receive/transmit package buffer sizes.

If a baud rate other than the default is to be used, then it is recommended that this is set during the power-up sequence using the Set Baud Rate command.

Recommended load-unload configuration

The Set Drive Configuration command provides access to several features that alter the tape drive’s behavior when loading or unloading cartridges. These give a large amount of flexibility in designing an automation controller.

HP’s experience suggests that certain configurations result in significantly fewer difficulties when integrating the HP Ultrium tape drive.

HP recommends configuration with the Auto-Eject feature disabled. If Auto-Eject is enabled, the drive will eject a cartridge in a variety of cases not directly controlled by the automation controller. These include receiving a SCSI (regardless of the method of instigating the load), completion of the image verification step when upgrading the tape drive’s micro-code using a firmware upgrade cartridge, and completion of a head-cleaning cycle when using a cleaning cartridge. These ejects can result in both the automation controller and the tape drive losing track of the location of the cartridge.

Supporting Ultrium features44

LOAD/UNLOAD command with the Load bit set to 0, various load failures

HP restricted

HP recommends configuration of the upgrade protect features to enabled. (the Upgrade Protect bit of the Set Drive Configuration command is set to 1). This will ensure that if a firmware upgrade cartridge is loaded inadvertently, the drive’s micro-code will not be upgraded unnecessarily.

If requested, HP will alter the default settings for Auto-Eject, Auto-Load, Auto-Thread, Clean Protect, and Upgrade Protect features in your particular variant of the firmware.

Recommended Get Drive Status polling frequency

HP recommends that the polling frequency of a Get Drive Status command should be in the range 2–5s, particularly during cartridge loading and unloading. This frequency should be sufficient to capture state changes in the drive while not adding significant processing overhead to the drive or automation controller.

ACI protocol communications retry

The ACI protocol specifies a comprehensive packet retry mechanism. Under certain timing conditions, especially for automation controllers that use a single microprocessor and multiplex the ACI communications from one tape drive to another, the automation controller can receive a response packet from the tape drive that it does not need. When this situation arises, the automation controller should send a positive acknowledgement control character, <ACK>, to the tape drive and discard the packet. Since the tape drive receives the <ACK>, it will not re-send the packet.

Upgrading the drive firmware

There are three methods of updating the firmware in the tape drive:

Firmware upgrade via tape

It is expected that firmware upgrades via tape will be done under the control of the library controller and the Operator Control Panel and independently of the host interface.

If the Upgrade Protect bit is set to 1 in the Set Drive Configuration command (which is recommended), the tape can be loaded into the drive in the usual manner, except that the ACI Load command must be sent to the drive and the Upgrade bit and Thread bit in byte 1 of the Load command must be set to 1.

If the Immediate Response bit in the ACI Load command is not set to 1 and the firmware upgrade failed (say due to an invalid image on the tape), the ACI Load command will report a

CONDITION

If the Immediate Response bit is set to 1 and the firmware upgrade fails, the automation controller can detect the failure by noting that the Tape Activity field in the Get Drive Status response returns to Idle and the tape drive does not enter its ACI initialization procedure.

• While the drive is preparing to upgrade the firmware, it will report Tape Activity = “Code

Update in Progress”.

• While it is actually upgrading the firmware, the drive will not respond to ACI commands.

• After the firmware upgrade has completed the drive will reset and send out an ENQ byte over

ACI.

After performing a firmware upgrade via tape it is recommended that the library controller checks that an ENQ byte is sent by the drive after it power-cycles at the end of the firmware upgrade

with appropriate sense key and additional sense.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 45

HP restricted

CHECK

process and that the normal power-up ACI command sequence is followed to ensure that the drive is configured correctly and to verify the firmware version and ACI version.

Firmware upgrade via the primary host interface

The library controller will not have direct visibility if a firmware upgrade of the tape drive is initiated via SCSI, hence it is recommended that the controller monitors for the that a firmware upgrade is taking place or has taken place.

• While the firmware image is being sent to the drive via SCSI, the drive responds to ACI

commands with status

• When the drive is actually upgrading the firmware, it will not respond to ACI commands.

• When the firmware download is complete, the drive will reset itself and send an ENQ control

character.

It is recommended that the same ACI command sequence be followed as if the drive had been power-cycled to ensure that the drive is configured correctly and to verify the firmware version and ACI protocol version.

BUSY.

Firmware upgrade via ACI

Two methods exist for updating firmware via ACI:

• Using the Send Firmware Image command. The automation controller sends the firmware image

in one data burst outside a normal packet.

• Using the Send Firmware Segment command. The automation controller sends the firmware

image in multiple packets.

HP intends to make the Send Firmware Image command obsolete in a future version of the ACI. Please use the Send Firmware Segment command in all new development. See the ACI specification for further details of both commands.

When the firmware download is complete, the drive will reset itself and send an ENQ control character. It is recommended that the same ACI command sequence be followed as if the drive had been power-cycled to ensure that the drive is configured correctly and to verify the firmware version and ACI version.

Library firmware upgrade via tape

The ACI specification allows for upgrading the automation controller firmware via tape. This functionality is not supported in current releases of drive firmware and will be added at a later date subject to customer needs.

Handling irregular cartridges

The purpose of this section is to indicate what can be seen over the ACI protocol if the host issues a

MOVE MEDIUM command to the library when an irregular cartridge (such as a cleaning cartridge,

expired cleaning cartridge, Ultrium 4 cartridge, or defective data cartridge) is in the storage element.

The following descriptions assume that the auto-eject bit in the Set Drive Configuration command has been set to 0 so that the cartridge will not be ejected from the drive unless an ACI Unload command is issued with the Eject bit set to 1.

Supporting Ultrium features46

HP restricted

Cleaning cartridge (HP-configured or Universal)

When a valid cleaning cartridge (one that has not expired) is loaded, behavior depends on the Clean Protect bit of the Set Drive Configuration command.

Clean Protect = 1 f the Clean Protect bit is set to 1, the drive will not thread the tape or clean the

drive until an ACI Load command with the Clean bit set to 1 is sent to the drive. If the Load command is sent without the Clean bit set the drive will return a

CHECK CONDITION. Also, if the “cleaning cartridge” is not in fact a cleaning

cartridge, the Load command with the Clean bit set to 1 will produce a

CONDITION

Clean Protect = 0 If the Clean Protect bit in the Set Drive Configuration command is set to 0, the

drive will thread the tape and clean the drive when a cleaning tape is loaded.

When the cleaning cartridge is seated in the drive, the ‘cartridge type’ field in the Get Drive Status RDATA will be set to 06h (cleaning).

While the drive is cleaning, the Cleaning bit in the Get Drive Status RDATA will be set to 1 and the Tape Activity field will be set to Ah (cleaning).

When cleaning has finished, if Auto-Eject is disabled, the cartridge will be in the ready eject position with the Cartridge Present, Write Protect, Ready Eject, and Ready Load bits set to 1, Cartridge Type = ‘Cleaning’, and Tape Activity = ‘Idle’. The cartridge can now be unloaded from the drive.

CHECK

Expired cleaning cartridge (HP-configured or Universal)

If an expired cleaning cartridge is loaded into the drive, the cartridge will be placed in the ready eject position with the Cartridge Present, Write Protect, Ready Eject, Ready Load, Media Error, TapeAlert, and Clean Expired bits set to 1, Cartridge Type = ‘Cleaning’, and Tape Activity = ‘Idle’. TapeAlert flag 22h will be set.

Non-HP Ultrium 1 cleaning cartridge

If a non-HP Ultrium 1 cleaning cartridge is loaded into the drive, the cartridge will not be recognized as a supported cartridge. The cartridge will be placed in the ready eject position with the Cartridge Present, Write Protect, Ready Eject, Ready Load, Media Error, and TapeAlert bits set to 1, Cartridge Type = ‘Unknown’, and Tape Activity = ‘Idle’. TapeAlert flag 17h will be set.

Unreadable data cartridge

If a data cartridge is loaded that cannot be read, the cartridge will be placed at the ready-to-eject position with the Cartridge Present, Write Protect, Ready Eject, Ready Load, Media Error, and TapeAlert bits set to 1, Cartridge Type = ‘Unknown’, and Tape Activity = ‘Idle’. TapeAlert flag 05h will be set.

Ultrium 5 and later data cartridges

If an Ultrium 5 or later data cartridge is loaded into the drive, the drive will recognize the cartridge as a non-supported cartridge. The cartridge will be placed at the ready-to-eject position with the Cartridge Present, Write Protect, Ready Eject, Ready Load, Media Error, and TapeAlert bits set to 1, Cartridge Type = ‘Unknown’, and Tape Activity = ‘Idle’. TapeAlert flag 0Ch will be set.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 47

HP restricted

Data cartridge with unreadable CM

If the Cartridge Memory cannot be read, the drive assumes that the cartridge is not supported. If the cartridge is loaded into the drive, it will be placed at the ready-to-eject point with the Cartridge Present, Write Protect, Ready Eject, Ready Load, Media Error, and TapeAlert bits set to 1, Cartridge Type = ‘Unknown’, and Tape Activity = ‘Idle’. TapeAlert flag 0Fh will be set.

Cartridge fails to seat or load

If a cartridge fails to seat or load, it will be placed at the ready-to-eject position with the Cartridge Present, Ready Eject, Ready Load, Media Error, and TapeAlert bits set to 1, Tape Activity = ‘Idle’. TapeAlert flag 04h will be set. If the cartridge type is recognized, this will be indicated in the Cartridge Type field, otherwise the field will indicate Cartridge Type = ‘Unknown’.

Cartridge cannot be loaded

HP recommends that GOOD status is not returned to the host for the MOVE MEDIUM command until the library controller has seen the Cartridge Load bit in the Get Drive Status RDATA set to 1. If the library controller does not see this bit set, HP recommends invoking an appropriate load re-try algorithm. After re-trying the load, if this bit is still not set to 1, assume that there is a problem with the cartridge. HP recommends that the library controller responds to the with

CHECK CONDITION, a sense key of Not Ready and additional sense of 5300 (media load or

eject failure), and then moves the cartridge back to the source element. If the Drive Error bit is set to 1 in the Get Drive Status RDATA, appropriate actions should be taken.

MOVE MEDIUM command

Valid firmware upgrade cartridge

If a firmware upgrade cartridge with a valid firmware image is loaded, and neither the library controller nor the host knows that the cartridge is a firmware upgrade cartridge, what occurs depends on the Upgrade Protect bit in the Set Drive Configuration command.

Upgrade Protect = 1 If the Upgrade Protect bit in the Set Drive Configuration command is set to 1,

it is assumed that the Upgrade bit in the Load command will be zero and no firmware upgrade will be performed. The cartridge will be placed at the ready-to-eject position with the Cartridge Present, Write Protect, Ready Eject, Ready Load, Media Error, and TapeAlert bits set to 1, Cartridge Type = ‘Firmware Upgrade’, and Tape Activity = ‘Idle’. TapeAlert flag 10h will be set.

Upgrade Protect = 0 If the Upgrade Protect bit in the Set Drive Configuration command is 0, a

firmware upgrade will be performed on the drive. While the drive is preparing to upgrade the firmware, it will report Tape Activity = “Code Update in Progress”. When actually upgrading the firmware the drive will not respond to ACI commands. After the firmware upgrade has completed the drive will reset and send out an ENQ byte over ACI.

It is recommended that the library controller follows the normal power-up ACI command sequence after receiving the ENQ byte to ensure that the drive is configured correctly and to verify the firmware version and ACI version.

Supporting Ultrium features48

HP restricted

Invalid firmware upgrade cartridge

If a firmware upgrade cartridge with an invalid firmware image is loaded, and neither the library controller nor the host knows that the cartridge is a firmware upgrade cartridge, again what occurs depends on the Upgrade Protect bit in the Set Drive Configuration command.

Upgrade Protect = 1 If the Upgrade Protect bit in the Set Drive Configuration command is set to 1,

Upgrade Protect = 0 If the Upgrade Protect bit in the Set Drive Configuration command is set to 0,

Frequently asked questions

it is assumed that the Upgrade bit in the Load command will be zero and no firmware upgrade will be performed on the drive. The cartridge will be placed at the ready-to-eject position with the Cartridge Present, Write Protect, Ready Eject, Ready Load, Media Error, and TapeAlert bits set to 1, Cartridge Type = ‘Firmware Upgrade’, and Tape Activity = ‘Idle’. TapeAlert flag 10h will be set.

the firmware upgrade process will start and the drive will thread the tape and read the image. During this time, the drive will report Tape Activity = “Code Update in Progress”. When the image has been read the drive will check whether the image is valid. As in this case the image is not valid, the drive will place the drive at the ready-to-eject position with the Cartridge Present, Write Protect, Ready Eject, Ready Load, Media Error, and TapeAlert bits set to 1, Cartridge Type = ‘Firmware Upgrade’, and Tape Activity = ‘Idle’. TapeAlert flags 10h and 22h will be set. The drive will not send out an ENQ byte and will not reset.

ACI protocol allows activities of the drive to be co-ordinated within a library. It provides several modes for operating HP Ultrium drives within tape libraries. In addition, the Cartridge Memory can, at the very minimum, provide an ‘electronic barcode’ facility to allow media tracking. HP is working with all the major tape library vendors to ensure that the full potential of these features are realized, and recognizes that ISV software support is a key part of this process. In advance of the release of the HP Ultrium Automation Cookbook, here are answers to some frequently asked questions:

Is there separate firmware for drives intended to go into libraries?

Specific variants of the firmware for automation drives enable the automation interface, which is not enabled in standalone drives. Otherwise the code is similar. A variant can only be downloaded if it matches the variant in the drive.

Cleaning

The tape drive tells the automation controller that a cleaning tape needs to be used through two bits in the ACI Get Drive Status command. The Cleaning Needed bit indicates deterioration in the write or read margin of the drive and hence it is recommended that a cleaning cartridge is inserted into the drive at the earliest opportunity. Following a successful clean, the Cleaning Needed bit will be cleared. The Cleaning Required bit indicates that the drive is unable to read or write unless the drive is first cleaned with a cleaning tape. It is recommended that a cleaning cartridge is loaded into the drive immediately. Following a successful clean, the Cleaning Required bit will be cleared.

Under normal circumstances, cleaning cartridges can be used for 50 cleanings.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 49

HP restricted

Resetting drives

The tape drive can be reset by the automation controller via the ACI Reset command or, in ACI mode, by pulling the ACI_RST_L line low (see “Rear Panel and Connectors”, Chapter 7 of the Hardware Integration Guide, Volume 1 of the HP LTO Ultrium Technical Reference Manual).

Resetting via the ACI Reset command

Two levels of reset via the Automation Interface are provided, namely ACI Reset and Drive Reset. The former resets the Automation Interface port and all SCSI parallel/Fibre Channel ports. The latter is equivalent to a power-on reset. Either reset method will interrupt the interface between the drive and host, with the Drive Reset potentially resulting in no End of Data written to tape.

It is therefore strongly recommended that an ACI reset command is not sent unless all other recovery methods have failed. Note that certain ACI commands (see Table 2 on page 43) can be queued behind outstanding SCSI commands giving the impression that the drive has stopped responding over the Automation Interface bus. (All command packages will be still be ack'ed even though the command will be queued.)

A SCSI interface reset will not affect the Automation interface.

NOTE: Following an upgrade of the drive firmware via either tape or SCSI, the drive will be reset

as if it had been powered up.

The implementation details are beyond the scope of this document currently.

Resetting using the ACI_RST_L line

The behavior when the ACI-RST_L line is activated depends on whether the drive is in ACI or ADI mode. In ACI mode activating this line will produce a power-on reset. Activating the ACI_RST_L in ADI mode will simply logout the ADT port.

Further details

• For more information about ACI, see “Automation Control Interface (ACI)” in Chapter 1, “Ultrium

Features”, of Background to Ultrium Drives, Volume 6 of the HP Ultrium Technical Manual.

• For more details of hardware integration, see Chapter 5, “Automating Drives”, of Hardware

Integration, Volume 1 of the HP Ultrium Technical Manual.

Backup software

You need backup application software that supports your Ultrium drive and tape library. For the latest list of appropriate backup packages, contact your tape library supplier.

Suitable backup applications include driver software that establishes the interface between the tape drive and the software. Applications usually recognize tape drives by their manufacturers’ ID string rather than their model number:

LTO 4 SCSI drive “HP Ultrium 4-SCSI”

Supporting Ultrium features50

HP restricted

LTO 4 SAS drive “HP Ultrium 4-SCSI” (“SCSI” is not a typo!) LTO 4 FC drive “HP Ultrium 4-SCSI” (“SCSI” is not a typo!)

Controlling data compression

The data compression hardware in HP Ultrium drives can detect whether incoming data is already compressed and will not attempt to compress it again. The drive can switch dynamically and automatically between compressing and non-compressing modes, thereby optimizing both compression ratio performance and data rate. As this mode of behavior is embedded in the Ultrium format, there is no need for host application software to switch the drive’s data compression on and off and HP strongly recommends to leave data compression at its default of 1 (on). This is set in the Select Data Compression Algorithm (SDCA) field of the Device Configuration mode page (page 10h).

Host applications may read the Data Compression mode page (0Fh) to determine that the drive is capable of data compression; the DCC bit is set to indicate this. HP recommends that the host does not attempt to modify either the Data Compression Enable (DCE) or Data Decompression Enable (DDE) bits.

Interpreting the current compression ratio and reporting it through applications is desirable for customers so that HP can accurately support customer installations. Without knowing the compression ratios achieved, it is difficult for support to suggest whether performance achieved is good or merely tolerable for customers’ data. HP recommends reporting compression and includes example source code for interpreting this functionality.

Other mode page information

Accessing Cartridge Memory without threading the tape

You can configure Ultrium drives so that when a cartridge is loaded, the Cartridge Memory can be accessed without threading the tape. This functionality is set with the Autoload field (byte 5, bits 0–2) in the Control mode page (0Ah). Zero (default) allows the cartridge to load in a conventional way. 001b or 010b allows the media to be loaded but not threaded, so the drive can read the Cartridge Memory contents. All other values for this field are reserved.

Buffer size at EW-EOM

The Buffer Size at EW-EOM field in the Device Configuration Mode page (10h) is set to zero. HP Ultrium tape drives automatically allow sufficient space between EW-EOM and “physical” EOT to satisfy backup applications.

Synchronize at EW-EOM

The SEW field (byte 10, bit3) on the Device Configuration mode page (10h) specifies how the drive behaves when reaching EW-EOM.

The parameter is fully supported. Default 0: the drive continues to buffer data objects after passing EW-EOM. If you require behavior compatible with earlier drive generations, send a command to set this bit to 1.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 51

HP restricted

MODE SELECT

Write delay time

The Write Delay Time field (bytes 6–7) on the Device Configuration mode page (10h) specifies the inactivity delay before the drive will automatically flush its data buffer. A value of 0 is an infinite delay; any other value is the delay in 100ms units.

This parameter is modifiable by a corresponds to 30s.

Rewind on reset

The Rewind on Reset field (byte 15, bits 3–4) on the Device Configuration mode page (10h) allows the host control over the action taken in the event of a bus reset.

If the field is 00 (default) or 10b the logical position is unchanged following a reset. If the field is set to 01b the media will be rewound to the logical beginning of media following a bus reset event.

Partition size

The Ultrium format only supports a single partition. The Medium Partition Mode page (11h) has no changeable fields. In the Device Configuration Mode page (10h), the Active Partition field should both be zero since multiple partitions are not supported.

NOTE: If the SET CAPACITY command is used to create logically shortened media, the partition size

is updated to reflect the new capacity.

In the Device Configuration mode page (10h) the Active Partition field must be 0 since multiple partitions are not supported.

MODE SELECT command. The default value is 12Ch (300d) which

Supporting Ultrium features52

HP restricted

6 Sense keys and codes

Sense keys—actions to take

Ultrium drives follow the ANSI definition of sense keys. The following table explains how the drives interpret sense key descriptions.

As sense keys and additional sense codes are intended to be hierarchical errors, the table recommends action for the host when a particular sense key is reported.

For more detailed recovery actions, see “Additional sense codes—actions to take” on

page 57.

Code Sense Key Interpretation

NO SENSE

RECOVERED ERROR

Action: The host will know what to do with this information, depending

These are informational/positional codes. The additional sense codes are not generally considered errors; they usually indicate some condition (such as hitting a filemark). The tape positional codes are mandatory for all sequential access devices. This use complies with SCSI-2, so it is generic.

For additional sense codes, see “0h—NO SENSE” on page 58.

upon the I/O operation at the time. Since the drive implements Progress Indication, this sense key may be returned if the drive is polled while an immediate operation is in progress.

CHECK CONDITION occurs with this sense key, and the additional

sense code is not recognized by the host, the software should just log the occurrence and continue. It will not be considered an error. The I/O should have completed without an error.

These errors have been recovered by the drive. The drive may report any type of recovered error additional sense code.

RECOVERED ERROR is returned is a MODE SELECT parameter is

truncated or if a TapeAlert event is being reported.

RECOVERED ERROR is the lowest priority sense key; it is only

reported when the I/O has completed with no other type of

CONDITION

For additional sense codes, see “1h—RECOVERED ERROR” on

page 59.

Console Message: None

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 53

CHECK

having occurred.

HP restricted

Code Sense Key Interpretation

NOT READY

This sense key generally means the host will have to wait for the drive to become

READY. Media access is not possible. See

“3h—MEDIUM ERROR” on page 62. Also see the Media Access

pre-execution check.

Action: The host needs to take one of the following actions:

• Wait until the drive becomes available.

• Issue some type of initializing command.

• Perhaps instruct the user to put the drive online.

• The handling of this sense key will depend upon the host’s

operating system and the additional sense code.

Console Message: Drive not ready - media access not possible

MEDIUM ERROR

This sense key indicates a failure that is probably due to a problem with the tape. The I/O did not complete. The I/O may have been partially attempted and data on the tape may have been altered. The drive will have retried an optimal number of times before reporting this sense key.

For additional sense codes, see “3h—MEDIUM ERROR” on

page 62.

Action: Recovery depends on the operating system or application. At the

very least, whatever the additional sense code, the software should log the error, terminate I/O to the drive, and pass the appropriate error to the calling application.

On read, the driver should discriminate between hard read failures caused by the media, and those resulting from an inability to decompress data.

Console Message: On write, ASQ 0C00h Write to tape failure

On write attribute, ASQ 1112h Write to CM failure On read, ASQ 70NNh Decompression

exception

On read attribute, ASQ 1112h Reading CM failure On read (otherwise) Read from tape failure On space Failed to locate

record on tape

Sense keys and codes54

HP restricted

Code Sense Key Interpretation

HW ERROR

This indicates that the current I/O operation has failed due to a hardware failure.The FRU code in the sense data should indicate which part of the hardware is bad. The drive should not be used again until corrective action has been taken. Specific recovery depends on the operating system and application.

For additional sense codes, see “4h—HW ERROR” on page 67.

Action: Whatever the additional sense code, the software should log the

error, terminate the I/O, and report the appropriate error to the calling application. Whether the drive requires any further corrective action before it can be used again (such as a reset or manual intervention) depends on the additional sense code.

Console Message: Tape drive hardware failure

ILLEGAL REQUEST

The last command sent to the drive, or the data sent because of the command, could not be accepted by the drive because it violated conditions imposed by the drive.

For additional sense codes, see “5h—ILLEGAL REQUEST” on

page 68. Also see the Illegal Command, Illegal Field, Fixed Bit,

Reservation and Parameter List pre-execution checks.

Action: The software can retry the I/O, or else it can terminate the I/O and

report an error to the calling application, particularly if the I/O has been retried a number of times and continues to fail with the same sense key. The specific retry or recovery strategy depends on the operating system.

Console Message: Illegal SCSI command requested

UNIT ATTENTION

The operating conditions of the drive have been changed in some manner that the host should be aware of. For example, the drive may have gone online or been reset, the Mode parameters may have been changed, a second host may have changed the drive’s operating conditions, and so on.

For additional sense codes, see “6h—UNIT ATTENTION” on

page 69. Also see the Unit Attention pre-execution check.

Action: Recovery depends on the device class and the additional sense

code. In general, the software should assume that mode settings and so on have been lost, and so should re-initialize the drive. The failed command will not have been executed and should be repeated.

Note that for parallel SCSI, the host driver may need to renegotiate transfer parameters (for example, by using a PPR message) before continuing.

Console Message: Tape drive operating conditions may have

changed

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 55

HP restricted

Code Sense Key Interpretation

DATA PROTECTION

This is an error if the I/O operation is attempting to access the media in some manner and failing because data on the media may not be accessed at this time (for example, because the tape is write-protected; or the drive is unable to decrypt data because the key is incorrect).

For additional sense codes, see “7h—DATA PROTECTION” on

page 72.

Also see the Media Write pre-execution check.

Action: Depends on the operating system. If this sense key occurs with other

than an additional sense code of 2700h, the software should log the error, terminate the I/O, and then send an error (operating system specific) to the calling application. Data on the tape will not have been altered.

Console Message: Media write-protected or encrypted

BLANK CHECK

An attempt was made to read unwritten media. Recovery from this sense key depends on the operating system. It may be regarded as an error because more data was expected by the host, or it may be an expected condition.

For additional sense codes, see “8h—BLANK CHECK” on

page 74.

Console Message: End-of-Data encountered

ABORTED

COMMAND

The drive has terminated the command. This could be caused by a problem related to the SCSI bus or Fibre Channel link. For example, it is reported if a target or LUN receives a second command from the same host before the previous command from that host has completed.

For additional sense codes, see “Bh—ABORTED COMMAND” on

page 75.

Action: Recovery depends on the additional sense code and the operating

system. In some cases, the host may want to retry the current I/O. If the additional sense code is 4E00h (overlapped commands attempted), the host may not want to retry the current I/O because the previous I/O will not have been completed.

Console Message: SCSI protocol problem

Sense keys and codes56

HP restricted

Code Sense Key Interpretation

VOLUME

OVERFLOW

Console Message: Physical End-of-Tape reached, unable to fit

Data could not be written because of a lack of remaining space on the tape. See the from this depends on the device class and the operating system. It is a “generic” sense key—the host should be able to recover from it without knowing the additional sense code.

For additional sense codes, see “Dh—VOLUME OVERFLOW” on

page 76.

remaining information on tape

WRITE and WRITE FILEMARK commands. Recovery

Additional sense codes—actions to take

These tables provide information about sense data, so that software can know which additional sense codes can be reported under which sense keys. It is important that the operating system makes all Request Sense data available to applications and, in interpreted form, to the end-user.

Actions are suggested for software to use when determining the recovery action for different sense keys and additional sense code and qualifying codes (ASC/ASCQ).

It is strongly recommended that the operating system and/or application use the entire ASC/ASCQ data to determine the appropriate recovery action.

The tables are in numerical order, not order of priority. That is, they do not suggest which sense keys should be checked first, nor do they recommend priorities for the devices to report errors.

NOTE: When the sense, additional sense code and qualifying sense keys are listed, the

software may look at all three keys to determine action. The drive should use that exact combination to report that particular error.

For example, a drive will report that it is not ready when there is no cartridge present by setting the sense key to command required). No other combination of sense key and additional sense may be used to report that particular condition.

NOT READY with additional sense of 0402h (LUN not ready, initializing

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 57

HP restricted

0h—NO SENSE

The following action applies to most additional sense codes in this group:

Action: For all additional sense codes except 82 82h, the action of the software depends on

the current I/O and what the operating system has been expecting. Recovery depends on the operating system. As a minimum, the software should pass an error to the calling application indicating the positional mark that has been encountered. The I/O can be retried if desired.

Code Meaning Comments

00 00 No additional sense The drive has no additional sense information for the

00 01 Filemark detected This indicates one of the following:

00 02 End of Tape detected A command completed early because End of Tape or the

host. The flags in the sense data indicate the reason for failure.

Action: see above

• A

READ or SPACE command was terminated early

because a filemark was encountered.

• Unsolicited Positional Sense has been set to indicate

“at a filemark”.

The Mark bit in the sense data will always be set.

Action: see above

physical end of the tape was encountered.

The EOM flag in the sense data will be set.

Action: see above

00 04 Beginning of Tape detected BOT was encountered during a space command. 00 16

00 18 00 19

00 1A

82 82 Drive requires cleaning The drive has detected that the heads need to be

Sense keys and codes58

Operation in progress Erase operation in progress Locate operation in progress Rewind operation in progress

Action: Either wait for the command to complete, or poll again

Action: Optionally, log the occurrence for information. It will not

The command is in progress and has not yet completed. This could be because another host initiated the command; or the command was sent in immediate mode. The Sense Key Specific Value field in the sense data will give some indication of how far the operation has progressed.

to see how it is progressing.

cleaned to maintain good operation.

be considered an error and the software will continue.

HP restricted

1h—RECOVERED ERROR

NOTE: Reporting of recovered errors defaults to OFF.

Action: In all cases, action depends on the device class and operating system.

Code Meaning Comments

37 00 Rounded parameter The drive needs to round off the value of a parameter

5D 00 Failure prediction threshold

exceeded

5D FF Failure prediction threshold

exceeded (false)

sent by degree of accuracy sent by the command.

Failure Prediction thresholds have been exceeded indicating that TapeAlert flags have been activated.

Action: Retrieve the TapeAlert log page to find out which flag

has been activated and advise the user accordingly. The Informational Exceptions Mode page has been sent

with the Test field set to 1 and the DExcpt field to 0, causing the drive to generate a false informational exception condition (a false device failure).

Action: Since the function of the Test field is simply to test that an

informational exception condition will produce a

CONDITION

the TapeAlert log, no action is necessary.

MODE SELECT because it cannot store it to the

CHECK

and that the exception will be reported to

2h—NOT READY

Code Meaning Comments

04 00 LUN not ready, no cause

to report

This is set if an unload is occurring with immediate report on, or initiated through the front panel, or a different host initiated the command. It is present for the duration of the unload or eject, after which the additional sense changes to 3A 00h (medium not present) or 0402h (logical unit not ready, initializing command required).

Action: 1. Issue a message to the console stating that the tape is

currently being unloaded from the drive.

2. Poll the drive until the additional sense changes to

3A 00h or 04 02h.

3. Instruct the user what to do, based on the application

and the previous sequence of commands.

4. Depending on the application, the software may

terminate the current I/O.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 59

HP restricted

Code Meaning Comments

04 01 LUN in process of

becoming ready

A media access command has been received while a load is occurring with immediate report on, or initiated through the front panel, or a different host initiated the command.

Action: 1. Effectively poll the drive by re-sending the command

until the media is loaded, when

UNIT ATTENTION with

additional sense of 28 00h will be set if the tape was inserted via the front panel. Otherwise poll the drive (TUR) until

GOOD status is reported. At this point the

command can be executed.

2. Report to the console that the drive is in the process

of loading tape.

04 02 LUN not ready, initializing

command required

Action: 1. The software must issue a

A cartridge is present in the drive but is not logically loaded.

LOAD before media access

is permitted.

2. Issue a message to the console to request a

user-initiated load, or to indicate drive status, or both.

04 07 Command in progress The tape drive is currently executing an immediate mode

command.

04 0C LUN not accessible, port in

unavailable state

The command is not available on the specified logical unit when sent to the drive through the specified port.

04 12 Logical unit offline The command cannot be executed because the specified

logical unit has not yet been configured via the appropriate port.

0B 01 Thermal limit exceeded There has been a failure due to the drive temperature

being outside the acceptable range.

Sense keys and codes60

HP restricted

Code Meaning Comments

30 03 Cleaning cartridge

installed.

A medium-access command has been sent to the drive while a cleaning cartridge was loaded.

Action: 1. Terminate the current I/O, and return the appropriate

error.

2. Send a message to the console indicating that a

cleaning cartridge is in the drive and a cleaning cycle is being performed.

3. Prompt the user to wait for the cartridge to be

ejected. In a library, the cartridge will be ejected when requested by the library or host. In an internal or external single drive, the cartridge will be ejected automatically.

4. Prompt the user to proceed with the next

application-specific activity.

5. Log the cleaning cycle in the system log.

3A 00 Medium not present A medium-access command has been received when no

cartridge is in the drive.

Action: As a minimum, issue a message to the console

indicating that a drive is present but no tape is loaded.

3A 04 Medium not present,

Media Auxiliary Memory accessible

A media access command has been received when the tape has been loaded but not threaded. This will be reported if the hold bit of the

LOAD CDB was set or the

Autoload field in the Control mode page is non zero.

3E 00 Logical unit has not

self-configured yet

This is set during power-up when it is not possible to send medium-access commands to the drive because mechanism tests are being executed. When the tests are complete, the additional sense changes to 3A 00h, 04 01h or 04 02h depending on whether a cartridge was present at power-on.

Action: 1. Issue a message to the console indicating that the

drive is powering up.

2. Effectively poll the drive until the drive transitions to

another state, at which point either execute the command or terminate the I/O.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 61

HP restricted

3h—MEDIUM ERROR

Code Meaning Comments

00 02 End of Tape detected A

0C 00 Write error The drive has failed to write data or filemarks to tape.

READ, SPACE, WRITE or WRITE FILEMARKS command

found EOT unexpectedly. This typically occurs when a drive cannot locate the target object on tape because the block count is too great.The EOM flag will be set.

Action: 1. Recovery action depends on the initiating action. As a

minimum, tell the calling application that physical EOP/M has been encountered. Also display this information as a console message.

2. Send any residue information to the calling

application.

This is probably due to bad media, but may be hardware-related. Residue information will normally be supplied.

Action: 1. Terminate the current I/O and return the appropriate

error.

2. The software should disable all further transactions to

the drive and mark the drive as ‘bad’.

3. The software should tell the user that a serious fault

has been detected with the drive and advise them to call their technical support.

4. Log the incident in the system log.

Sense keys and codes62

HP restricted

Code Meaning Comments

11 00 Unrecovered read error A read from tape has failed. This is probably due to bad

media, but may be hardware-related.

Action: 1. Terminate the current I/O and return the appropriate

error.

2. Send a console message that an unrecovered error on

write has occurred.

3. Determine whether the error is deferred, and report

the last successful operation and the failed operation to the calling application.

4. Log the error and all recovery actions in the system

log.

Recovery action is as follows:

1. Use Log Sense to find the age and state of the tape

and the drive. Based on this, ask the user to clean the drive or replace the tape.

2. If the fault is drive-related, ask the user to retry the

operation after the drive has been cleaned.

3. If the fault is with the media, prompt the user to back

up the data to a new tape, restart the application and discard the current tape.

11 12 Media Auxiliary Memory

read error

An error has occurred while attempting to write to MAM. The cartridge should not be used for further backups but should be able to be used for restoring data.

14 00 Recorded entity not found A

SPACE or LOCATE command failed because of the drive

could not find the target of the operation because of a format violation.

Action: 1. Terminate the I/O and return the appropriate error.

2. Send a message to the console indicating that EOD

could not be found because the tape has a corrupt format.

3. Prompt the user to back up the data to another tape

and discard the current one.

4. Log the incident in the system log.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 63

HP restricted

Code Meaning Comments

14 03 End of data not found A read-type operation failed because a format violation

related to a missing EOD data set, or there was an attempt to read a brand new tape.

Action: 1. Terminate the I/O and return the appropriate error.

2. Send a message to the console indicating that EOD

could not be found because the tape has a corrupt format.

3. Prompt the user to back up the data to another tape

and discard the current one.

4. Log the incident in the system log.

30 00 Incompatible medium

installed

A write-type operation could not be executed because it is not supported on the tape type that is loaded.

Action: 1. Terminate the I/O and return the appropriate error.

2. Send a message to the console indicating that writing

is not allowed on the type of tape that is currently loaded.

3. Prompt the user to insert a different tape type.

4. Log the incident in the system log.

5. The calling application can retry the operation.

30 01 Cannot read media,

unknown format

LOCATE or SPACE command has found the tape is in a

format not supported by the drive.

Action: 1. Terminate the current I/O, and return the appropriate

error.

2. Send a message to the console indicating that the

tape is in a format not supported by the drive.

3. Prompt the user to eject the cartridge and insert a

valid one.

4. Log the incident in the system log.

30 02 Cannot read media,

incompatible format

READ command could not be completed because the

logical format is not correct.

Action: 1. Terminate the current I/O, and return the appropriate

error.

2. Send a message to the console indicating that the

tape is wrongly formatted.

3. Prompt the user to eject the cartridge and insert a

valid one.

4. Log the incident in the system log.

30 04 Cannot write medium The tape’s Cartridge Memory is bad so that the tape is

unusable.

Sense keys and codes64

HP restricted

Code Meaning Comments

30 07 Cleaning failure A cleaning operation was attempted but could not be

completed for some reason.

Action: Use another cleaning cartridge because the current one

has expired.

30 0D WORM medium—integrity

check failed

The drive has detected an inconsistency when performing an integrity check on a WORM cartridge. The cartridge may have been tampered with. Data can be read from the cartridge by setting the WTRE bit in the Device Configuration mode page. The event should be logged for audit purposes.

Action: 1. Terminate the current I/O, and return the appropriate

error.

2. Send a message to the console indicating that the

drive has detected an inconsistency in the WORM cartridge that indicates it may have been tampered with.

3. Alert the user that the cartridge may have been

tampered with.

4. Log the incident in the system log for audit purposes.

5. Initiate WTRE-controlled EOD with clear warnings to

the end-user that the data may have been tampered with.

31 00 Medium format corrupted

READ or SPACE has tried to read data that is in a format

that is recognized but which is not valid.

Action: 1. Terminate the current I/O, and return the appropriate

error.

2. Send a message to the console indicating that there is

a problem with the format of the tape in the drive.

3. Prompt the user to eject the media and insert a valid

Ultrium tape.

4. Log the incident in the system log.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 65

HP restricted

Code Meaning Comments

3B 00 Sequential positioning

error

The drive has failed to read data off tape. There are two possibilities:

• The current command (such as

WRITE) failed to complete successfully.

READ, SPACE, REWIND,

• The logical position has been lost.

Action: 1. Attempt to recover by executing a

REWIND command

to return to a known position such as BOT.

2. Space to the position of the last known successful

command and retry the failing command.

3. If this is unsuccessful, terminate the current I/O, and

return the appropriate error.

4. Prompt the user to back up the data to a new

cartridge and discard the old one.

5. Log the incident in the system log.

50 00 Write append error A write-type command failed because the point at which

to append data was unreadable. This was probably caused by a powerfail, or SCSI bus or Fibre Channel link reset, preventing the drive from completing a write operation properly and appending an EOD.

Action: 1. Terminate the current I/O and return the appropriate

error.

2. Tell the user that the append point is unreadable

3. Advise the user to back up the data to new media and

reformat the failing tape.

4. Log the incident in the system log.

52 00 Cartridge fault A command could not be completed because of a fault

with the tape cartridge.

Action: 1. Terminate the current I/O and return the appropriate

error.

2. Tell the user that a serious fault has been detected

with the tape cartridge.

3. Advise the user to discard this cartridge and select a

new one.

4. Log the incident in the system log.

Sense keys and codes66

HP restricted

Code Meaning Comments

53 00 Media load or eject failed A load or eject has failed.

53 04 Medium thread or

unthread failure

4h—HW ERROR

The following actions apply to most additional sense codes for HW ERROR sense key:

Action: 1. Terminate the current I/O and return the appropriate

error.

2. Inform the user that a serious fault has been detected

with the tape cartridge.

3. Advise the user to discard this cartridge and select a

new one.

4. Log the incident in the system log.

The threading or unthreading operation failed.

Action: 1. Terminate the current I/O and return the appropriate

error.

2. Inform the user that a serious fault has been detected

with the tape cartridge.

3. Advise the user to discard this cartridge and select a

new one.

4. Log the incident in the system log.

Action: 1. Terminate the current I/O and return the appropriate error.

2. The software should disable all further transactions to the drive and mark the

drive as ‘bad’.

3. The software should tell the user that a serious fault has been detected with the

drive and advise them to call their technical support.

4. Log the incident in the system log.

Code Meaning Comments 40 XX Diagnostic failure on

component XX

Action: see above

44 00 Internal target failure This code is used to report hardware and firmware

Action: see above

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 67

A self-test command has detected an error, or a command is prohibited from execution due to failure of a previous diagnostic. “XX” is a vendor-specific code indicating the failing component.

related hard errors that occur when the drive encounters an “impossible” situation.

HP restricted

Code Meaning Comments

53 01 Unload tape failure The tape unload failed because it cannot be physically

82 83 Bad microcode detected The data transferred to the drive during a firmware

5h—ILLEGAL REQUEST

The following actions apply to all additional sense codes in this group:

Action: 1. Terminate the current I/O and return the appropriate error.

2. Log the incident in the system log.

3. Print out all the

identify the location of the offending bits or bytes.

4. Refer to the table of Drive Error Codes in Volume 3, SCSI Interface, of the

Ultrium Technical Reference Manual to understand why the current values were rejected.

5. This is likely to be an application fault. Send a message to the console “Illegal

SCSI request to tape drive”.

completed at this point in time.

Action: see above

upgrade is corrupt or incompatible with the drive hardware.

Action: see above

REQUEST SENSE data bytes and check bytes 15 through 17 to

Code Meaning Comments

1A 00 Parameter list length error A

20 00 Invalid command opcode The drive does not recognize the opcode of the

24 00 Invalid field in CDB The drive has detected an invalid field in a command

25 00 LUN not supported The command was addressed to a logical unit number

26 00 Invalid field in parameter

list

26 01 Parameter not supported A request for an invalid page number has been sent.

Sense keys and codes68

MODE SELECT parameter list sent to the drive contains

one of the following:

• An incomplete Mode Parameter header (must be

4bytes)

• An incomplete Mode Block Descriptor (must be 0 or

8bytes)

• An incomplete Mode page

command it has received.

descriptor block.

that does not exist. The drive detected an invalid field among the command

parameters sent during the data phase.

HP restricted

Code Meaning Comments

26 04 Invalid release of persistent

reservation

The Persistent Reservation holder has tried to release the persistent reservation using the

PERSISTENT RESERVE OUT

command, but the Scope or Type supplied was invalid.

2C 00 Command sequence

invalid

The sequence of SCSI commands is invalid.

Example 1: The use of the echo buffer was invalid. A

WRITE BUFFER command is necessary before a READ BUFFER

command.

Example 2: Another initiator has already started a firmware download process.

2C 0B Not reserved If the “Only If Reserved” bit is set in the Device

Configuration Mode page, and the drive does not hold a (persistent) reservation, some commands will not be allowed to execute. In other words, some commands can only be executed if the drive is reserved.

3B 0C Position past BOM A

SET CAPACITY command was received when the

logical position was not BOT, a necessary condition for this command.

53 02 Medium removal prevented An unload operation failed to eject the tape because

medium removal has been prevented.

55 03 Insufficient resources A buffer has reached its full capacity. 55 04 Insufficient registration

resources

FC interface only: There is only space for requests from 32 initiators to register, using

PERSISTENT RESERVE OUT

commands.

55 06 Media Auxiliary Memory

full

There is insufficient space in the Host Attribute area in MAM to fit the attribute that need to be written.

Action: Check MAM attribute 0004h (MAM Space Remaining)

to identify how much space remains in MAM.

6h—UNIT ATTENTION

Code Meaning Comments

28 00 Not ready to ready

transition

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 69

A tape has been loaded successfully into the drive and is now ready to be accessed.

Action: 1. The host should be polling, receiving a

CONDITION

with sense key 2h (NOT READY) and additional sense of 04 01h (LUN in process of becoming ready), and expecting this transition.

2. Send the console message: “Tape drive in process of

becoming ready”.

3. After the transition, send the console message “Tape

loaded - media may have changed”.

HP restricted

CHECK

Code Meaning Comments

29 01 Power-on reset The drive has powered on since the host last accessed it.

Action: 1. The action of the calling application depends on the

current I/O and what the operating system is expecting.

2. For parallel SCSI, the host should renegotiate transfer

parameters, and reconfigure the drive with any host-specific operating parameters (burst size, bus activity limit, fixed or variable mode, and so on).

3. The host should then report to the console that the

drive has been reset.

4. The I/O can be retried if desired.

29 02 SCSI bus reset SCSI drives: The drive has received a SCSI reset signal

since the host last accessed it.

FC drives: The drive has received its first process login.

The drive will be implicitly logged out after a Target Reset so this ASCQ will be posted after the host has performed port/process login.

Action: As for 29 01h

29 03 Bus device reset SCSI drives: The drive has received a SCSI bus device

reset message since the host last accessed it.

FC drives: The drive has received a process login when it

was previously logged in for a particular host.

Action: As for 29 01h

29 04 Internal firmware reboot The drive has reset itself.

Action: The host may renegotiate transfer parameters.

29 05 Transceivers to SE

SCSI drives only: The transceivers on the bus have been

reset to Single-Ended.

Action: As for 29 01h

29 06 Transceivers to LVD

SCSI drives only: The transceivers on the bus have been

reset to LVD.

Action: As for 29 01h

29 07 I_T nexus loss occurred The drive has lost the connection with the initiator (host

server).

Action: The host may renegotiate transfer parameters.

2A 01 Mode parameters changed The Mode parameters for the drive have been changed

by a host other than the one issuing the command.

ATTENTI ON

is set for all hosts following a MODE SELECT

UNIT

command, apart from the host that issued the command. This code will only be returned in a multi-host environment.

Sense keys and codes70

HP restricted

Code Meaning Comments

Action: When operating the drive in this type of environment, the

following actions should occur:

1. The calling application receiving this code should

issue a

MODE SENSE command requesting the drive

to return all parameters.

2. The application should check those parameters over

which it has configuration control, to ensure that the current configuration of the drive does not conflict with what the application expects.

3. If it finds discrepancies, the application can either

reconfigure the drive to the original values, or halt and report an error.

4. If an error is reported, a console message must be

displayed, and information logged to the system log.

2A 02 Log parameters changed The Log parameters for the drive have been changed by

an initiator other than the one issuing the command.

2A 03 Reservations pre-empted A

PERSISTENT RESERVE OUT command with the Clear

service action removed all reservations and the persistent reservation.

2A 04 Reservations released A

PERSISTENT RESERVE OUT command executed. The

original persistent reservation has been replaced with another of a different type or removed completely.

2A 05 Registrations pre-empted A

PERSISTENT RESERVE OUT command was executed

which removed all registrations.

2A 11 Data encryption parameters

changed by another

The encryption parameters that this initiator was using have been modified by another initiator.

initiator

2A 12 Data encryption parameters

changed by vendor specific event

The encryption parameters that this initiator was using have been modified because of a vendor specific event (such as tape unload or reservation released).

3F 01 Firmware upgraded The firmware in the drive has just been changed by a

WRITE BUFFER or MAINTENANCE OUT command,

Firmware Update Cartridge, ADI or ACI.

3F 05 Device identifier changed A

SET DEVICE IDENTIFIER command has been successful.

3F 0E Reported LUNs data has

changed

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 71

HP restricted

7h—DATA PROTECTION

Code Meaning Comments

26 10 Data decryption key fail limit

reached

27 00 Write-protected This is set if a write operation (

2A 13 Data encryption Key Instance

Counter has changed

30 00 Incompatible medium

installed

30 05 Cannot write medium,

incompatible format

SECURITY PROTOCOL OUT command has failed because

an incorrect key has been sent to the drive followed by a read, and this has happened ten times consecutively.

Action: Further

SECURITY PROTOCOL OUT commands will not

succeed until one of these actions has been taken:

• Unload and reload the current tape.

• Power-cycle the drive.

Advise the user that they are using the wrong key, or that there is a possible security attack.

WRITE, WRITE FILEMARKS, or

ERASE) is requested for a write-protected cartridge.

Action: 1. Terminate the current I/O and return the appropriate

error.

2. Send a message to the console indicating that the drive

has been trying to write to a write-protected tape.

3. Subsequent action depends on the application.

The drive received a

WRITE command from an initiator that

had locked its encryption parameters to a specific Key Instance Counter, whose value has now changed.

Action: 1. Issue

SECURITY PROTOCOL IN command to find out what

the current encryption parameters are.

2. Issue a

SECURITY PROTOCOL OUT command to set the

correct security parameters.

3. Notify the user of a possible system configuration issue,

since several initiators are trying to control encryption parameters.

A write-type operation could not be executed because it is not supported on the tape type that is loaded.

Action: 1. Terminate the I/O and return the appropriate error.

2. Send a message to the console indicating that writing

is not allowed on the type of tape that is currently loaded.

3. Prompt the user to insert a different tape type.

4. Log the incident in the system log.

5. The calling application can retry the operation.

For example, the drive tried to write to a tape of an incompatible generation.

Sense keys and codes72

HP restricted

Code Meaning Comments

30 0C WORM—overwrite attempted A write operation could not be executed because an

overwrite has been attempted on a WORM cartridge. This may be because an overwrite backup was specified instead of an appended backup.

Action: 1. Terminate the I/O and return the appropriate error.

2. Send a message to the console indicating that an

attempt has been made to overwrite on a WORM cartridge.

3. Prompt the user to either use a new cartridge or change

the operation to an appended backup.

4. Log the incident in the system log.

30 0D WORM—integrity check

failed

An attempt has been made to write to a WORM cartridge when the drive has detected inconsistencies while checking its integrity.

Action: 1. Terminate the I/O and return the appropriate error.

2. Send a message to the console indicating that the drive

has detected an inconsistency with the WORM cartridge that indicates that it may have been tampered with.

3. Advise the user that the tape may have been tampered

with and they should use a different cartridge.

4. Log the incident in the system log for audit purposes.

74 0 0 Security error Generic security (encryption/decryption) error. For

example, the Decryption mode is set to RAW and the drive has been asked to read, but the hardware does not allow this Decryption mode on the current block.

74 0 1 Unable to decrypt data The drive encountered encrypted data while reading, but

decryption mode is not enabled.

Action: Send a

SECURITY PROTOCOL OUT command with

Decryption Mode set to Decrypt or Mixed, and the Key field specifying the correct decryption key.

74 0 2 Unencrypted data

encountered while decrypting

Action: Send a

The decryption mode is enabled but the drive encountered non-encrypted data while reading.

SECURITY PROTOCOL OUT command with

Decryption Mode set to Disable; alternatively with Decryption Mode set to Mixed and the Key field specifying the correct decryption key.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 73

HP restricted

Code Meaning Comments

74 03 Incorrect data encryption key The drive read into a block of data which could not be

74 0 4 Cryptographic integrity

validation failed

74 0 5 Key-associated data

descriptors changed

8h—BLANK CHECK

decrypted with the current decryption key.

Action: 1. Send a

SECURITY PROTOCOL IN with the page set to

Next Block Encryption Status to obtain information about the next block on tape.

2. Send a

SECURITY PROTOCOL OUT command with

Decryption Mode set to Decrypt or Mixed, and the Key field specifying the correct decryption key

The next block failed the integrity validation process while the drive was attempting to read it.

Action: Data may have been compromised, so the tape should not

be trusted. Log the incident in the system log and notify the user.

The Key-associated data descriptor values have changed compared to the values in the last recorded read.

Action: None. The application may continue reading but it is

advisable to send a

SECURITY PROTOCOL IN command with

the page set to Next Block Encryption Status to find out information about the next block on tape. The application may decide to change its own security parameters.

Code Meaning Comments

00 05 End of Data (EOD)

detected

A encountered EOD.

Action: 1. Terminate the current I/O and return the appropriate

2. Send a console message saying that EOD has been

3. Recovery depends on the calling application and

Sense keys and codes74

READ or SPACE command terminated early because it

error to the calling application indicating that EOD has been encountered.

encountered.

what was expected.

HP restricted

Code Meaning Comments

14 03 End of Data not found A read-type operation failed because a format violation

Action: 1. Terminate the current I/O and return the appropriate

Bh—ABORTED COMMAND

The following action applies to all codes in this group:

Action: 1. Terminate the current I/O and return the appropriate error.

2. Log the incident in the system log.

3. Send the console message “SCSI command aborted - low-level failure on SCSI

bus” or for a Fibre Channel drive, “SCSI command aborted - low-level failure on Fibre Channel link”.

related to a missing EOD data set. The most likely cause is a tape with corrupt format (perhaps from a powerfail when the tape was being written).

error.

2. Send a message to the console indicating that EOD

could not be found because the tape has corrupt format.

3. Prompt the user to back up the data to another tape

and discard the current one.

4. Log the incident in the system log.

Code Meaning Comments

3F 0F Echo buffer overwritten A

47 00 SCSI parity error The drive has detected a parity error, for example during

47 01 Data phase CRC error

detected

47 03 Information Unit CRC error

detected

48 00 Initiator Detected Error

message received

4B 00 Data phase error ACI received more raw data than expected.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 75

READ BUFFER command has been received with Echo

Buffer mode set, and the echo buffer has been overwritten by a different host from that which issued the

READ BUFFER command.

Message phase or Data phase. For parallel SCSI only in non-Information Units mode, the

drive has detected a CRC error during the data out phase.

For parallel SCSI only in Information Units mode, the drive has detected a CRC error during the command IU or data out phase.

An Initiator Detected Error message has been received, and the previous phase (the phase in which ATN was asserted) was invalid.

HP restricted

Code Meaning Comments

4E 00 Overlapped commands

attempted

74 0 8 Digital signature validation

failure

Dh—VOLUME OVERFLOW

Code Meaning Comments

00 02 End of Tape detected A

NOTE: When the sense, additional sense and qualifying sense keys are listed, the software may

look at all three keys to determine action. The drive should use that exact combination to report that particular error.

For example, a drive will report that it is not ready when there is no cartridge present by setting sense key = additional/qualifying sense keys may be used to report that particular condition.

NOT READY with additional/qualifying sense keys = 04 02. No other combination of

A host has selected the drive even though it already has a command outstanding.

The SCSI command used to download the new firmware image failed because the firmware image digital signature could not be correctly validated.

WRITE or WRITE FILEMARKS command has encountered

EOT or the physical end of tape. The EOM flag will be set.

Sense keys and codes76

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:

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

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.

Allow the user to try recovery.

For any fault, the system allows the user to use simple recovery commands such as

LOAD/UNLOAD.

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 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.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 77

HP restricted

LOGICAL UNIT RESET Task

REWIND or

5. Parallel SCSI only:

SCSI Bus Reset.

If the

LOGICAL UNIT RESET Task Management function fails to clear the problem, the System

Supervisor or technical support person should be able to perform a controlled hard SCSI Bus Reset as follows:

• Lock other users out cleanly.

• Go to a minimal-system single-user mode.

• Close all applications.

•Execute a hard SCSI Bus Reset. The process should

Allow the System Supervisor or support person to perform a controlled hard

not require a complete shutdown and reboot of the system.

Monitoring the condition of the drive and media

Through TapeAlert, an HP LTO Ultrium tape drive constantly monitors the condition of the mechanism and media, and presents the results in a form that host software can readily use and users can easily understand.

You can find a general description of TapeAlert in the diagnostics section of the User Guide.

Supporting TapeAlert

HP has refined the TapeAlert standard to include the Predictive Failure flag. If the drive sets this flag, the host should display a message that the drive is “about to fail and should be replaced”. There should be no impact on ISV software if the TapeAlert standard is already fully supported.

NOTE: Reading the TapeAlert log clears all the flags, as defined in the TapeAlert specification.

The TapeAlert Log page consists of 64 flags that indicate potential problems with the drive, and that allow host software to suggest appropriate corrective action to the user. For example, if Flag 20 (“Clean Now”) is set, the software should advise the user to clean the tape heads.

The drive maintains both SSC and ADC TapeAlert flags because the two standards state different clearing conditions.

• The drive maintains separate copies of the SSC TapeAlert flags for each port. These are only

cleared when a port retrieves its TapeAlert flags with a LOG SENSE command with page code 2Eh, provided the TAPLSD (TapeAlert Prevent LOG SENSE Deactivation) mode parameter is zero in the Device Configuration Extension mode page.

• The drive also maintains a set of ADC TapeAlert flags, which are shared by each port. The ADC

TapeAlert flags are not cleared when a port retrieves them with a LOG SENSE command with page code 12h.

See Chapter 4 of Volume 3, SCSI Interface, of the HP Ultrium Technical Reference Guide for details of the TapeAlert log page.

Exception handling78

HP restricted

Flags

The following table lists the flags that could potentially be supported in tape drives. Of these, flags 3,4,5,6,20,22 and 31 are mandatory for drives such as Ultrium drives that support cleaning cartridges.

The flags are grouped into the following sections:

• Flags 1 to 19: For tape drive write/read management

• Flags 20 to 25: For cleaning management

• Flags 26 to 39: For tape drive hardware errors

• Flags 50 to 60: For additional tape drive errors

For each flag, the message that the host software should display to the user is given, together with the cause of the flag being set. The Type column classifies the flags by seriousness into the following three groups:

I Information A suggestion to the user.

W Warning The user is advised to take action. Performance or data

may be at risk otherwise.

C Critical! Take action immediately.

The Set column indicates if the flag can be set by LTO 4 drives. The other flags are supported but never set.

Flag Type Set Recommended Host Message Cause

Flags for Tape Drive Write/Read Management

1 Read warning W

2 Write warning W The tape drive is having problems

3 Hard error W

The tape drive is having problems



reading data. No data has been lost, but there has been a reduction in the capacity of the tape.

writing data. No data has been lost, but there has been a reduction in the capacity of the tape.

 The operation has stopped because

an error has occurred while reading or writing data which the drive cannot correct.

The drive is having severe trouble reading.

The drive is having severe trouble writing.

This flag is set for any unrecoverable read/write/positioning error, and is cleared internally when the tape is ejected. The flag is set as an explanation of the error in conjunction with one of the recovery action flags 4, 5, or 6.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 79

HP restricted

Flag Type Set Recommended Host Message Cause

4 Media C  Your data is at risk:

1. Copy any data you require from

this tape.

2. Do not use this tape again.

3. Restart the operation with a

different tape.

5 Read failure C The tape is damaged or the drive is

faulty. Call the tape supplier’s helpline.

6 Write failure C

 The tape is from a faulty batch or the

tape drive is faulty:

Media performance is severely degraded or the tape can no longer be written or read.

This flag is set for any unrecoverable read/write/positioning error caused by faulty media. It is cleared internally when the media is ejected.

The drive can no longer read data from the tape.

The flag is set for any unrecoverable read error where the diagnosis is uncertain and could either be a faulty tape or faulty drive hardware. It is cleared internally when the tape is ejected.

The drive can no longer write data to the tape.

1. Use a good tape to test the drive.

2. If the problem persists, call the

tape drive supplier’s helpline.

7 Media life W The tape cartridge has reached the

end of its calculated useful life:

1. Copy any data you need to

another tape.

2. Discard the old tape.

8 Not data grade W Not relevant to Ultrium drives. 9 Write-protect C

Exception handling80

 You are trying to write to a

write-protected cartridge.

Remove the write-protection or use another tape.

HP restricted

The flag is set for any unrecoverable write/positioning error where the diagnosis is uncertain and could either be a faulty tape or faulty drive hardware. It is cleared internally when the tape is ejected.

The media has exceeded its specified life.

A write command was attempted to a write-protected tape.

Flag Type Set Recommended Host Message Cause

10 No removal I  You cannot eject the cartridge

because the tape drive is in use.

Wait until the operation is complete before ejecting the cartridge.

11 Cleaning media I

12 Unsupported

format

13 Recoverable

mechanical cartridge failure

14 Unrecoverable

mechanical cartridge failure

15 Memory chip in

cartridge failure

16 Forced eject C

17 Read-only

format

18 Tape directory

corrupted on load

 The tape in the drive is a cleaning

cartridge.

 You have tried to load a cartridge of

a type that is not supported by this drive.

 The operation has failed because the

tape in the drive has experienced a mechanical failure:

1. Discard the old tape.

2. Restart the operation with a

different tape.

C The operation has failed because the

tape in the drive has experienced a mechanical failure:

1. Do not attempt to extract the tape

cartridge.

2. Call the tape drive supplier’s

helpline.

 The memory in the tape cartridge has

failed, which reduces performance. Do not use the cartridge for further write operations.

 The operation has failed because the

tape cartridge was manually de-mounted while the tape drive was actively writing or reading.

C You have loaded a cartridge of a

type that is read-only in this drive. The cartridge will appear as write-protected.

 The tape directory on the cartridge

has been corrupted. File search performance will be degraded. The tape directory can be rebuilt by reading all the data on the cartridge.

A manual or software unload was attempted when Prevent Medium Removal was in force.

A cleaning cartridge is loaded in the drive.

Attempted load of an unsupported tape format.

The tape has snapped or suffered some other mechanical failure in the drive, but the tape can still be ejected.

The tape has snapped or suffered some other mechanical failure in the drive and the tape cannot be ejected.

The LTO-CM chip has failed in cartridge.

A manual or forced eject occurred while the drive was writing or reading.

A write command has been attempted to a tape whose format is read-only in this drive.

The drive was powered down with a tape loaded, or a permanent error prevented the tape directory being updated.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 81

HP restricted

Flag Type Set Recommended Host Message Cause

19 Nearing media

life

Flags for Cleaning Management

20 Clean now C

21 Clean periodic W The tape drive is due for routine

22 Expired

cleaning media

I  The tape cartridge is nearing the end

of its calculated life. It is recommended that you:

1. Use another tape cartridge for

your next backup.

2. Store this tape cartridge in a safe

place in case you need to restore data from it.

 The tape drive needs cleaning:

If the operation has stopped, eject the tape and clean the drive.

• If the operation has not stopped,

wait for it to finish and then clean the drive.

• Check the tape drive user’s

manual for cleaning instructions.

cleaning:

1. Wait for the current operation to

finish.

2. Use a cleaning cartridge.

3. Check the tape drive user’s

manual for cleaning instructions.

 The last cleaning cartridge used in

the tape drive has worn out:

The tape may have exceeded its specified number of passes.

The tape drive has detected that it needs cleaning. The flag is cleared internally when the drive is cleaned successfully.

The drive is ready for a periodic cleaning.

The cleaning tape has expired.

23 Invalid cleaning

cartridge

Exception handling82

1. Discard the worn-out cleaning

cartridge.

2. Wait for the current operation to

finish.

3. Use a new cleaning cartridge.

 The last cleaning cartridge used in

the tape drive was an invalid type:

1. Do not use this cleaning cartridge

in this drive.

2. Wait for the current operation to

finish.

3. Use a valid cleaning cartridge.

HP restricted

The flag is set when the tape drive detects a cleaning cycle was attempted but was not successful. It is cleared internally when the next cleaning cycle is attempted.

An invalid cleaning tape type was used.

Flag Type Set Recommended Host Message Cause

24 Retension

requested

25 Dual-port

interface error

Flags for Tape Drive Hardware Errors 26 Cooling fan

failure

27 Power supply

failure

28 Power

consumption

29 Drive

maintenance

W The tape drive has requested a

retension operation.

W A redundant interface port on the

tape drive has failed.

W A tape drive cooling fan has failed. A fan inside the drive

W A redundant power supply has failed

inside the tape drive enclosure. Check the enclosure user’s manual for instructions on replacing the failed power supply.

W The tape drive power consumption is

outside the specified range.

W Preventive maintenance of the tape

drive is required.

The drive is having trouble reading or writing that will be resolved by a retension cycle.

One of the interface ports in a dual-port configuration (in other words, Fibre Channel) has failed.

mechanism or enclosure has failed.

A redundant PSU has failed inside the tape drive enclosure or rack subsystem.

The tape drive power consumption is outside the specified range.

The drive requires preventive maintenance (not cleaning).

30 Hardware A C

31 Hardware B C

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 83

Check the tape drive user’s manual for preventive maintenance tasks or call the tape drive supplier’s helpline.

 The tape drive has a hardware fault:

1. Eject the tape or magazine.

2. Reset the drive.

3. Restart the operation.

 The tape drive has a hardware fault:

1. Turn the tape drive off and then

on again.

2. Restart the operation.

3. If the problem persists, call the

tape drive supplier’s helpline.

HP restricted

The drive has a hardware fault from which it can recover through a reset.

The drive has a hardware fault that is not read/write related or that it can recover from through a power cycle.

The flag is set when the tape drive fails its internal power-on self-tests. It is not cleared internally until the drive is powered off.

Flag Type Set Recommended Host Message Cause

32 Interface W The tape drive has a problem with

the application client interface:

1. Check the cables and cable

connections.

2. Restart the operation.

33 Eject media C

34 Download fail W

35 Drive humidity W Environmental conditions inside the

36 Drive

temperature

37 Drive voltage W The voltage supply to the tape drive

38 Predictive failure C A hardware failure of the drive is

39 Diagnostics

required

Flags for Additional Tape Drive Errors

50 Lost statistics W

 The operation has failed:

1. Eject the tape or magazine.

2. Insert the tape or magazine

again.

3. Restart the operation.

 The firmware download has failed

because you have tried to use the incorrect firmware for this tape drive.

Obtain the correct firmware and try again.

tape drive are outside the specified humidity range.

 Environmental conditions inside the

tape drive are outside the specified temperature range.

is outside the specified range.

predicted. Call the tape drive supplier ‘s helpline.

W The tape drive may have a hardware

fault. Run extended diagnostics to verify and diagnose the problem.

Check the tape drive user’s manual for instructions on running extended diagnostic tests.

Flags 40–49 are not currently used.

 Media statistics have been lost at

some time in the past.

The drive has identified an interface fault.

Error recovery action.

Firmware download failed.

The drive’s humidity limits have been exceeded.

The drive is experiencing a cooling problem.

Drive voltage limits have been exceeded.

Failure of the drive’s hardware is predicted.

The drive may have a hardware fault that may be identified by extended diagnostics (using a SEND DIAGNOSTIC command).

The drive or library has been powered on with a tape loaded.

Exception handling84

HP restricted

Flag Type Set Recommended Host Message Cause

51 Tape directory

invalid at unload

52 Tape system

area write failure

53 Tape system

area read failure

54 No start of data C The start of data could not be found

55 Loading failure C

W  The tape directory on the tape

cartridge just unloaded has been corrupted. File search performance will be degraded. The tape directory can be rebuilt by reading all the data.

C The tape just unloaded could not

write its system area successfully:

1. Copy the data to another tape

cartridge.

2. Discard the old cartridge.

C The tape system area could not be

read successfully at load time.

Copy the data to another tape cartridge.

on the tape:

1. Check that you are using the

correct format tape.

2. Discard the tape or return the

tape to your supplier.

 The operation has failed because the

media cannot be loaded and threaded.

An error has occurred preventing the tape directory being updated on unload.

Write errors occurred while writing the system area on unload.

Read errors occurred while reading the system area on load.

The tape has been damaged, bulk erased, or is of an incorrect format.

The drive is unable to load the cassette and thread the tape.

56 Unrecoverable

load failure

1. Remove the cartridge, inspect it

as specified in the product manual, and retry the operation.

2. If the problem persists, call the

tape drive supplier’s help line.

 The operation has failed because the

tape cannot be unloaded:

1. Do not attempt to extract the tape

cartridge.

2. Call the tape driver supplier’s

help line.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 85

HP restricted

The drive is unable to unload the tape.

Flag Type Set Recommended Host Message Cause

57 Automation

interface failure

58 Firmware failure W

59 WORM

medium—integr ity check failed

60 WORM

medium— overwrite attempted

C The tape drive has a problem with

the automation interface:

1. Check the power to the

automation system.

2. Check the cables and cable

connections.

3. Call the supplier’s helpline if the

problem persists.

 The tape drive has reset itself due to

a detected firmware fault. If the problem persists, call the supplier’s helpline.

 The tape drive has detected an

inconsistency while checking the WORM tape for integrity. Someone may have tampered with the cartridge.

 An attempt has been made to

overwrite user data on a WORM tape:

1. If you used a WORM tape

inadvertently, replace it with a normal data tape.

2. If you used a WORM tape

intentionally, check that:

• the software application is

compatible with the WORM tape format you are using.

• the cartridge is bar-coded

correctly for WORM.

Flags 61–64 are not currently used

The drive has identified a fault in the automation interface.

There is a firmware bug.

Someone has tampered with the WORM tape.

The application software does not recognize the tape as WORM.

Note that often messages will not appear in isolation. For example, message 01h (“The tape drive is having problems reading data.”) is likely to appear with a message suggesting remedial action, such as message 04h (“You are advised to copy any data...”) or message 14h (Clean Now).

Each flag is cleared to zero in the following circumstances:

• At power-on.

• When specified corrective action has been taken, such as using a cleaning cartridge.

• When the TapeAlert Log page is read.

Exception handling86

HP restricted

NOTE: Once cleared, a flag cannot be set again until the specified clearing conditions are met.

So, for example, if the cartridge in the drive is not of data grade, once flag 8 has been cleared, it cannot be set again until the cartridge has been removed.

Designing software to use the TapeAlert log

When writing software to take advantage of the ability of a drive to predict problems and actions that a user should take, it is important not to exclude drives that do not support this feature. For this reason, the application should first check whether the TapeAlert Log page is supported by the drive and then use one of two methods to access the information:

• Use the

This means that the tape drive reports or more TapeAlert flags are set. When

MODE SELECT Informational Exceptions mode page to enable “Check Condition” mode.

CHECK CONDITION on the next SCSI command after one

CHECK CONDITION is received, the host software should

behave as follows: a. It issues a

CONDITION CONDITION

b. The software reads the TapeAlert log page to discover which flags are set (even for

CONDITION

Note that when

CHECK CONDITION is not in error and will have completed successfully. It follows that the

REQUEST SENSE command. Additional sense of 5D00h indicates that the CHECK

was caused by TapeAlert. This enables the software to distinguish CHECK s caused by actual errors and those resulting from a TapeAlert flag being set.

CHECK

s caused by actual errors).

CHECK CONDITION results from TapeAlert, the command that reported the

software should not repeat the command.

• Read the TapeAlert log page using

• Immediately after a SCSI

LOG SENSE at the following times:

CHECK CONDITION/REQUEST SENSE cycle.

• At the end of each tape where a backup or restore spans multiple tapes. The host must read the TapeAlert log page before a tape is ejected.

• At the end of a backup or restore.

It is also advisable to poll the TapeAlert log page every 60 seconds while the tape drive is idle.

Using “Check Condition” mode is recommended over polling because it guarantees that the software will be aware of any TapeAlert flag being set. It is theoretically possible that TapeAlert information could be missed when polling, though the higher the frequency of polling, the smaller the chance is.

TapeAlert models

SSC-3rev02 specifies three optional methods for an application to monitor activation of TapeAlert flags. LTO 4 drives support two of those models:

• Polling the TapeAlert log page (2Eh) or the TapeAlert Response log page (12h);

• Configuring the device server to establish an Informational Exception condition (via mode page

1Ch) upon activation of one or more TapeAlert flags.

These are described in more detail below.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 87

HP restricted

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.

Exception handling88

HP restricted

5. Optionally, automate the recommended recovery actions if there are multiple tape drives or

autoloaders present. For example, the application could perform a cleaning cycle in response to flags 20 (Clean

Now) and 21 (Clean Periodic). It could perform a tape copy for flags 4 (Media performance degraded) and 7 (Media life expired), and then retire the suspect tape cartridge.

This provides an opportunity for applications to add value to the TapeAlert capability of the drives.

NOTE: An application must not fail a backup job as a result of TapeAlert information. It should use

the information as a preventative measure, taking action to avoid failure, or encouraging the user to take action. It should also retain the log information to help in diagnosis if a job does eventually fail.

One-Button Disaster Recovery (OBDR)

NOTE: FC drives and drives in libraries do not support OBDR.

All HP Ultrium parallel-SCSI products support HP’s One-Button Disaster Recovery (OBDR) technology. This provides the fastest possible, one-step approach to regenerating a single server without using additional floppy disks or CD-ROMs.

For a general overview, see “OBDR and CD-ROM Emulation” in Chapter 1, “Ultrium Features”, of Background to Ultrium Drives, Volume 6 of the HP Ultrium Technical Manual. For details of the SCSI implementation, see “CD-ROM Emulation” in Chapter 1, “Interface Implementation”, of SCSI Interface, Volume 3 of the HP Ultrium Technical Manual.

For details of how to use OBDR see the appropriate User’s Guide.

To identify whether the firmware supports OBDR, look for the string “$DR-10” in bytes 43–48 of the Inquiry data.

Supporting OBDR

The OBDR implementation in HP Ultrium drives is functionally identical to that in DAT/DDS, so no additional design or coding should be required in order to support it. The only effort needed should be in software testing.

Responding to the ‘Clean’ LED

NOTE: HP recommends that software applications use the TapeAlert log, which should mean that

potential tape or cleaning problems are flagged and corrected before the drive ever reaches the point of displaying the ‘Clean’ LED.

If during normal operation, the drive detects an excessive number of RWW retries, the ‘Clean’ LED is lit. If this happens, a user should follow this procedure:

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 89

HP restricted

1. Clean the heads and try the operation again.

2. If the ‘Clean’ LED is lit again, repeat the operation with another tape cartridge. If this clears the

‘Clean’ LED, it indicates that the original cartridge is at fault. Copy the data from the cartridge onto a new one and discard the old cartridge.

The ‘Clean’ LED is cleared by a cleaning cycle.

Providing pass-through mode

It is important for Drivers and Logical Device Managers to provide a pass-through mode that can return information or commands specific to a vendor’s product. The need for this is two-fold:

• Systems must support a great variety of new devices.

• All tape drives are similar to a degree; Drivers and Logical Device Managers tend to provide

connectivity based on the assumption that 80% of all SCSI tape drives behave identically.

Pass-through mode offers the following advantages:

• Peripheral manufacturer can provide value-added diagnostics and support applications over

and above those that may be shipped with a system or application.

• System supervisors and operators can take advantage of specific product features otherwise

excluded because the driver or manager only caters for 80% of SCSI drives.

• Technical support people have access to low-level device-specific information likely to be

unavailable otherwise.

Requirements for drivers and logical device managers

Drivers and Logical Device Managers must allow the user to do the following:

1. Create either a 6-byte, 10-byte, 12-byte or 16-byte SCSI Command Descriptor Block.

2. Allocate a write buffer or file for any data associated with the SCSI command that will be sent to

the drive.

Allocate a read buffer or file for any data associated with the SCSI command that will be returned by the drive.

3. Link the command and data buffers.

4. Launch the command.

5. The driver should use its standard

whether the command completed successfully or not. The caller must have access to the raw

REQUEST SENSE data.

6. View any returned data.

Exception handling90

CHECK CONDITION and REQUEST SENSE routines to report

HP restricted

Glossary

ANSI American National Standards Institute, which sets standards for, amongst other things,

SCSI and the safety of electrical devices.

BOM Beginning Of Media. The first point on the tape that can be accessed by the drive.

buffered mode A mode of data transfer in write operations that facilitates tape streaming. It is selected

by setting the Buffered Mode Field to 1 or 2 in the SCSI header.

compression A procedure in which data is transformed by the removal of redundant information in

order to reduce the number of bits required to represent the data.

compression ratio A measure of how much compression has occurred, defined as the ratio of the amount

of uncompressed data to the amount of compressed data into which it is transformed. The LTO-DC algorithm can typically achieve a compression ratio of between 2:1 and 4:1 depending on the nature of the data.

decompression A procedure in which the original data is generated from compressed data.

ECMA European Computer Manufacturers Association. The European equivalent of ANSI.

EOD End Of Data. An area that signifies the end of the valid data. If new data is written

over a larger quantity of old data, it is possible for data to exist after EOD, but because it is after EOD, this old data is no longer valid.

MODE SELECT Parameter List

EOM End Of Media format. The last usable point on the tape.

EW-EOM Early Warning End Of Media. A physical mark or a device-computed position on the

tape that tells the drive that it is approaching EOM.

filemark A mark written by the host. It does not necessarily separate files; it is up to the host to

assign a meaning to the mark.

FRU Field Replaceable Unit, an assembly or group of components that is replaced in its

entirety by Service Engineers when it contains a fault.

hard error An uncorrectable data error.

host The host computer system acting as controller for the drive.

load The process in which the drive takes in an inserted cartridge and goes online.

LUN Logical Unit Number, by which different devices at a particular SCSI ID can be

addressed individually. The drive has a fixed LUN of 0.

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 91

HP restricted

offline The drive is offline if the tape is currently unloaded or not in the drive. The host has

limited access, and cannot perform any commands that would cause tape motion. The host can, however, load a tape, if one is inserted, and can execute any diagnostic tests that do not require tape motion.

online The drive is online when a tape is loaded. The host has access to all command

operations, including those that access the tape, set configurations and run diagnostic tests.

RWW see ”read-while-write”

read-while-write RWW improves data integrity by reading data immediately after it is written and

writing the data again if an error is found.

TapeAlert A set of 64 flags is held in the TapeAlert log that indicate faults or predicted faults with

the drive or the media. By reading this log, host software can inform the user of existing or impending conditions, and can, for example, advise the user to change the tape.

WORM Write Once, Read Many

HP restricted

Index

ACI 17, 35, 37

cleaning 49 command set 38 commands that affect streaming 38 firmware upgrade 46 protocol communications retry 45 reserved fields 43 resetting drives 50

supporting 42 ACI 4.3, new features 39 additional sense codes 57 ADI 35 ANSI 10, 91 autochangers 17 Automation Control Interface see ACI

backup applications 11 backward compatibility 39 barcode support, WORM 25 block mode

fixed-length 11

variable-length 11 block size, maximum 28 BOM 91 buffer

maximizing use 11

size at EW-EOM 51 buffered mode 91

capacity

of tape 20

remaining 20

using SET CAPACITY 21 Cartridge Memory 12, 20

responding to tape data 25 cartridge memory 19, 35

unreadable 48

using instead of headers 28

cartridges

identifying types 19 irregular 46 unreadable 47

WORM 22 Clean LED 89 cleaning 49, 58, 90 cleaning cartridges 47, 61 cleaning tape heads 12 commands, ACI 38 commands, non-immediate 11 compression 91

controlling 51

ratio 91 configuration 15

data compression 91

controlling 11, 51 data transfer size 11 decompression 91 design goals 13 Diagnostic logs 13 diagnostics, failure 67 displaying drive information 13 documents, related 7 drive

checking integrity 13

detecting speed 27

displaying information 13

initialization 15 drive technology family 16 drivers 15

ECMA 10, 91 eject, failed 67 encryption support 41 EOD 91

detected 74

not found 64, 75

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 93

HP restricted

EOM 58, 62, 91 EOPD 22 EOT 76 errors

hard 91 parity 75 read 63

write 62, 66 escalation procedure 77 EW-EOM 91 exception handling 77

faults, predicting 13, 78 filemarks 91

detected 58

use of 32 firmware revision 13 firmware upgrade 45, 48

loading an invalid cartridge 49 fixed-length block mode 11 flags, TapeAlert 79 format, corrupt 65 front panel LEDs 89 FRU 91

generation, finding 16

hard error 91 host 91

identifying tape cartridge types 19 INCITS 10 initialization 15 INQUIRY command 13 inquiry string recovery 15 integrity of WORM media 24 irregular cartridges 46 ISO 10

LEDs, Clean 89 LEDs, Use Cleaning Tape 89

libraries 49 library controller, slave 37 load

count 25

failed 67 loading 91 load-unload configuration 44 LOG SENSE data 20 logs

Diagnostic 13

performance 28

supporting 29

Tape Capacity 20

TapeAlert 13, 25, 78, 92 LTO-CM 19, 35

using instead of headers 28 LUNs 91

not ready 59

support for additional 17

MAM

full 69

write error 63 media labels 23 memory, cartridge 19 mode

fixed-length block 11

pass-through 90

variable-length block 11 mode parameters 70 monitoring condition 13, 78 monitoring tape use 12

NetWare drivers 15 new features in ACI 4.3 39 non-immediate commands 11

OBDR 89 offline 92 one-button disaster recovery 89 online 92 operating system drivers 15 optimizing performance 11, 27

HP restricted

parity error 75 partition size 52 pass-through mode 38, 90 performance factors 27

drive-related 31 format-related 32

host-related 29 performance log 28 polling frequency, Get Drive Status 45 power-up sequence 43 problems, predicting 13, 78 product ID 15 product revision level 15, 16

read errors 63 reading the TapeAlert log 88 remaining capacity 20 reserved fields 43 reset 50 revision level 15, 16 RWW 92

retries 89

retry counts 26

SCSI pass-through mode 38 sense keys

ABORTED COMMAND 56, 75

BLANK CHECK 56, 74

DATA PROTECTION 56, 72

HW ERROR 55, 67

ILLEGAL REQUEST 55, 68

MEDIUM ERROR 54, 62

NO SENSE 53, 58

NOT READY 54, 59

RECOVERED ERROR 53, 59

UNIT ATTENTION 55, 69

VOLUME OVERFLOW 57, 76 SET CAPACITY command 21 slave to a library controller 37 speed, detecting 27 status of tape 20 support for additional LUNs 17

Tape Capacity log 20 tape heads, cleaning 12 tape integrity 24 TapeAlert

informational exception usage model 88 polling usage model 88

TapeAlert log 13, 25, 78, 92

reading 88

using 87 tapemarks, use of 32 tapes

capacity 20

cleaning cartridge 61

faulty 66

identifying 28

identifying types 19

monitoring use 12

status 20

use of 12

using 19

WORM 22 tests, read/write 13 time-out values 28 transfer size 11, 27

effect on performance 32 troubleshooting 13

Ultrium

finding the format generation 16

supporting features 35 UNIX drivers 15 upgrading firmware 45 Use Cleaning Tape LED 89

variable-length block mode 11 vendor ID 15

Windows drivers 15 WORM 92 WORM media 22 write errors 62, 66 write-protect 72

HP LTO Ultrium 4 drives technical reference manual, volume 2: software integration 95

HP restricted

HP LTO 4 SAS, LTO 4 SCSI, LTO 4 FC User Manual

Specifications and Main Features

Frequently Asked Questions

User Manual

Contents

Related documents

Documents specific to HP Ultrium drives

Documentation map

Installation and configuration

Operation

Cartridges

Interface

Maintenance and troubleshooting

Dealing with errors

LTO Ultrium features

General documents and standardization

1 Designing backup applications

Optimizing performance

Large data transfer size

Data compression control

Non-immediate commands

Managing the use of tapes

Information in Cartridge Memory

Cleaning tape heads

Monitoring tape use

TapeAlert

Diagnostic logs

Displaying drive information

Drive tests

Design goals for LTO backup applications

2 Configuration and initialization

Operating system drivers

Inquiry string recovery

Support for additional LUN

Enabling additional LUN support

Supporting additional LUNs

LTO cartridge memory

Identifying tape cartridge types

Tape status and capacity

Finding the remaining capacity

Interpreting Log Sense data

Using the SET CAPACITY command

WORM media

How WORM media works

Changes to SCSI commands

Re-writing media labels

Allow overwrite of last filemarks before the EOD data set

Using CM to check tape integrity

Behavior with a missing or inconsistent EOD value

Unique media identifier

Barcode support

Responding to Cartridge Memory data

Load count

RWW retry counts

4 Factors affecting performance

Ways of optimizing performance

Detecting the drive’s speed

Ensuring the recommended minimum transfer sizes

Media type identification

Using Cartridge Memory instead of tape headers

Using the Performance Log page for diagnosing problems

Time-out values

Recommended support of log pages

Factors affecting performance

Host-related factors

Drive-related factors

Format-related factors

5 Supporting Ultrium features

Cartridge Memory (LTO-CM)

Automation interface

Automation/Device Interface (ADI)

Automation Control Interface (ACI)

Modes of usage through ACI

ACI command set

ACI commands that affect drive streaming performance

New features in ACI 4.3

Further details

Supporting the ACI protocol

Recommended ACI time-out values

Treatment of reserved fields