Software - Oracle Documentation

Software - Oracle Documentation
software
integration
hp ultrium drives
technical reference manual
generation 2 SCSI and FC drives
volume 2: software integration guide
Part Number: C7379–90900 Volume 2
Edition 2, February 2003
<Bold Header>
Notice
This document contains proprietary
information which is protected by
copyright. All rights reserved. No
part of this document may be
photocopied, reproduced or
translated to another language
without the prior written consent of
Hewlett-Packard.
The information contained in this
document is subject to change
without notice.
Hewlett-Packard 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 direct, indirect,
special, incidental or consequential
damages in connection with the
furnishing, performance, or use of
this document.
© Copyright 2003 by
Hewlett-Packard Limited
Revision History
Version
Date
Edition 1
Edition 2
Changes
All. HTML Generation 1 version.
Feb 2003
All. PDF Generation 2 SCSI and FC drive version.
This document is frequently revised and updated. To find out if there is a
later version, please ask your HP OEM Representative.
2
<Bold Header>
The Purpose of this Manual
This is one of five volumes that document HP Ultrium drives. This volume
provides background information for driver and application developers. The
following products are covered. Capacities are when the drive is using data
compression with a compression ratio of 2:1, where applicable:
3 HP Ultrium Generation 2 Full-Height SCSI Internal Drive
3 HP Ultrium Generation 2 Full-Height FC Internal Drive
Note
Throughout this manual frequent reference is made to SCSI
commands. For more information on SCSI commands for
HP Ultrium drives see volume 3, The SCSI Interface, of the
HP Ultrium Technical Reference Manual set. Ordering details
are given below.
Related Documents
The following documents provide additional information:
Documents Specific to HP Ultrium Drives
3 Hardware Integration Guide, volume 1 of the HP Ultrium Technical
Reference Manual
3 The SCSI Interface, volume 3 of the HP Ultrium Technical Reference
Manual
3 Specifications, volume 4 of the HP Ultrium Technical Reference Manual
3 HP Ultrium Configuration Guide, volume 5 of the HP Ultrium Technical
Reference Manual
3 Background to Ultrium Drives, volume 6 of the HP Ultrium Technical
Reference Manual
Please contact your HP supplier for copies.
3 The features and benefits of HP Ultrium drives are discussed in the
HP Ultrium Technology White Paper.
3 For a general backgrounder on LTO technology and licensing, go to
http://www.lto-technology.com .
3
Documentation Map
The following will help you locate information in the 6-volume Technical
Reference Manual:
Drives—general
SCSI Drives
Connectors
Controller architecture
Front Panel LEDs
Mechanism and hardware
Specifications
FC Drives
1 HW Integration: ch. 7 1 HW Integration: ch. 4
6 Background: ch. 4
1 HW Integration: ch. 6 1 HW Integration: ch. 3
6 Background: ch. 3
4 Specs
Installation and Configuration
SCSI Drives
FC Drives
1 HW Integration: ch. 7 1 HW Integration: ch. 4
2 SW Integration: ch. 2 2 SW Integration: ch. 2
Determining the configuration
1
External drives (SCSI only)
HW Integration: ch. 5
n/a
1 HW Integration: ch. 1
In Libraries
1 HW Integration: ch. 4
n/a
In Servers (SCSI only)
1 HW Integration: ch. 3
In Tape Arrays (SCSI only)
n/a
1 HW Integration: ch. 8
n/a
Modes of Usage (SCSI only)
n/a
Optimizing performance (SCSI only) 1 HW Integration: ch. 8
2 SW Integration: ch. 4
5 UNIX Config
UNIX configuration
Connectors
Operation
SCSI Drives
External drives (SCSI only)
In Libraries
In Servers (SCSI only)
In Tape Arrays (SCSI only)
4
FC Drives
1 HW Integration: ch. 5
n/a
1 HW Integration: ch. 1
1 HW Integration: ch. 4
n/a
1 HW Integration: ch. 3
n/a
Cartridges
SCSI Drives
Cartridge Memory (LTO-CM)
Cartridges
Features
Managing the use of cartridges
Use of cartridges
FC Drives
2 SW Integration: ch. 5
6 HW Integration: ch. 5
1 HW Integration: ch. 9 1 HW Integration: ch. 5
6 HW Integration: ch. 5
2 SW Integration: ch. 1
2 SW Integration: ch. 3
Interface
SCSI Drives
FC Drives
SCSI Guide
3 SCSI
Commands
3 SCSI: ch. 4
1 HW Integration: ch. 10 1 HW Integration: ch. 6
3 SCSI: ch. 1
2 SW Integration: ch. 3
3 SCSI: ch. 2
3 SCSI: ch. 4
Error codes
Implementation
Interpreting sense data
Messages
Mode pages
—see the MODE SENSE command
3 SCSI: ch. 3
Pre-execution checks
2 SW Integration: ch. 6
3 SCSI: ch. 4
Responding to Sense Keys and ASC/Q
Sense Keys and ASC/Q
—see REQUEST SENSE command
Maintenance and Troubleshooting
SCSI Drives
Cleaning
FC Drives
2 SW Integration: ch. 5
2 SW Integration: ch. 7
1 HW Integration: ch. 5
n/a
1 HW Integration: ch. 1
In Libraries
1
In Servers (SCSI only)
HW Integration: ch. 4
n/a
1 HW Integration: ch. 3
n/a
In Tape Arrays (SCSI only)
2 SW Integration: ch. 7
Monitoring drive and tape condition
2 SW Integration: ch. 1
Software troubleshooting techniques
External drives (SCSI only)
5
Dealing with Errors
SCSI Drives
FC Drives
1 HW Integration: ch. 10 1 HW Integration: ch. 6
2 SW Integration: ch. 5
Handling errors
6 Background: ch. 4
How error correction works
3 SCSI: ch. 4
Logs—see the LOG SENSE command
2 SW Integration: ch. 7
Recovering from write and read errors
2 SW Integration: ch. 3
Software response to error correction
2 SW Integration: ch. 3
Software response to logs
2 SW Integration: ch. 7
TapeAlert log
Error Codes
Ultrium Features
SCSI Drives
Adaptive Tape Speed (ATS)
6 Background: ch. 1
Autoload
1 HW Integration: ch. 2
Automation Control Interface (ACI)
1 HW Integration: ch. 2
6 Background: ch. 1
Cartridge Memory (LTO-CM)s
1 HW Integration: ch. 2
2 SW Integration: ch. 5
6 HW Integration: ch. 5
Data Compression, how it works
6 Background: ch. 5
2 SW Integration: ch. 5
6 Background: ch. 1
6 Background: ch. 1
2 SW Integration: ch. 7
Data Compression, managing
Design principles
OBDR and CD-ROM emulation
Performance optimization
Performance, factors affecting
Software design
Supporting Ultrium features
Ultrium Format
6
FC Drives
1 HW Integration: ch. 8
n/a
2 SW Integration: ch. 1
2 SW Integration: ch. 4
2 SW Integration: ch. 1
2 SW Integration: ch. 5
6 Background: ch. 2
General Documents and Standardization
3 Small Computer System Interface (SCSI-1), ANSI X3.131-1986. This is the
ANSI authorized standard for SCSI implementation, available through
ANSI
3 Enhanced Small Computer System Interface (SCSI-2), ANSI X3T9.2-1993
Rev. 10L, available through ANSI
Copies of General Documents can be obtained from:
ANSI 11 West 42nd Street
New York, NY 10036-8002
USA
ISO CP 56
CH-1211 Geneva 20
Switzerland
ECMA 114 Rue du Rhône
CH-1204 Geneva
Switzerland
Global Engineering Documents 2805 McGaw
Irvine, CA 92714
USA
Tel: +41 22 849 6000
Web URL: http://www.ecma.ch
Tel: 800 854 7179 or 714 261 1455
7
8
contents
Contents
The Purpose of this Manual 3
Related Documents 3
Documents Specific to HP Ultrium Drives 3
Documentation Map 4
General Documents and Standardization 7
1
Designing Backup Applications 13
Optimizing Performance 13
Large Data Transfer Size 13
Data Compression Control 14
Non-Immediate Commands 14
Managing the Use of Tapes 14
The Tape Log 15
Cleaning Tape Heads 15
Monitoring Tape Use 15
TapeAlert 16
Diagnostic Logs 16
Displaying Drive Information 16
Drive Tests 17
Online Help 17
Design Goals for LTO Backup Applications
2
17
Configuration and Initialization 19
Operating System Drivers 19
Inquiry String Recovery 19
Product ID, first 8 bytes 20
Product ID, last 8 bytes 20
Product Revision Level 21
Contents
9
Example 21
Support for Additional LUN 22
Enabling Additional LUN Support
Supporting Additional LUNs 22
3
22
Use of Tapes 23
LTO Cartridge Memory 23
Tape Status and Capacity 23
Finding the Remaining Capacity 24
Interpreting Log Sense Data 24
Responding to Tape Log Data 26
Load Count 26
RAW Retry Counts 26
4
Factors Affecting Performance 29
Ways of Optimizing Performance 29
Detecting the Drive’s Speed 30
Ensuring the Recommended Minimum Transfer Sizes 30
Media Type Identification 30
Using Cartridge Memory Instead of Tape Headers 30
Tuning Performance Using the Performance Log Page 31
Time-Out Values 31
Recommended Support of Log Pages 32
Factors Affecting Performance 32
Host-Related Factors 32
Drive-Related Factors 34
Format-Related Factors 36
5
Supporting Ultrium Features 37
Cartridge Memory (LTO-CM) 37
Cleaning 37
Resetting Drives 38
Backup Software 39
Drive Model ID String 39
Controlling Data Compression 39
Support of Data Compression 39
10
Contents
Other Mode Page Information 40
Buffer Size at EW-EOM 40
Partition Size 40
Sense Keys and Codes 41
Sense Keys—Actions to Take 41
Additional Sense Codes—Actions to Take
0h—NO SENSE 46
1h—RECOVERED ERROR 46
2h—NOT READY 47
3h—MEDIUM ERROR 49
4h—HW ERROR 53
5h—ILLEGAL REQUEST 54
6h—UNIT ATTENTION 56
7h—DATA PROTECTION 58
8h—BLANK CHECK 58
Bh—ABORTED COMMAND 59
Dh—VOLUME OVERFLOW 60
7
Contents
6
45
Exception Handling 61
Typical Escalation Procedure 61
Monitoring the Condition of the Drive and Media 62
Supporting TapeAlert 63
Monitoring the Condition of the Drive and Media 63
Flags 63
Flags for Tape Autoloader Errors 70
Designing Software to Use the TapeAlert Log 72
Reading the TapeAlert Log 73
Responding to the ‘Clean’ LED 74
Providing Pass-Through Mode 75
Requirements for Drivers and Logical Device Managers
Recovering from Write and Read Errors 76
Write Recovery 76
Read Recovery 76
Reading Through Media Errors 76
75
Contents
11
Glossary 79
Index 83
12
Contents
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:
3 Use large data transfer sizes.
3 Control and monitor data compression.
3 Ensure directory information is safe and accurate.
3 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 29.
Large Data Transfer Size
Applications should use large data transfer sizes to make better of the Ultrium
drive’s internal buffers (64 MB). A good goal to set is at least 64 kilobytes
each for read or write operation:
3 For fixed-length block mode reads and writes, provided the block size
times the number of blocks to be transferred is at least 64 kilobytes, drives
will provide peak performance. Small block sizes (512 bytes) are
Designing Backup Applications
13
acceptable so long as they are written and read in fixed-length block
mode in large transfers.
3 For variable-length block mode reads and writes, the transfer length
should be at least 64 kilobytes.
Data Compression Control
Ultrium drives have built-in hardware data compression. Backup applications
should incorporate features to:
3 Control the compression capability of the drive.
3 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 39.
Non-Immediate Commands
Performance can be improved by only using immediate Write Filemarks
commands.
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 23 for more information.
14
Designing Backup Applications
The Tape Log
The LTO Cartridge Memory holds the System Area for each Ultrium tape that
contains data about the tape’s history. This Tape log can be used to calculate
the error-rate history for the tape, which in turn helps to determine the
probability of a successful backup. This information can be used to warn
against backing up on a tape of dubious quality.
The Tape log also shows how many times a tape has been loaded or
unloaded which gives an indication of the condition of the tape. The
application can then warn a user to discard the tape and use a new one, if
necessary.
The ‘Clean’ LED on the front of HP Ultrium drives indicates when a cleaning
cartridge should be used. This information can be retrieved by a SCSI
REQUEST SENSE command looking at the CLN bit in the sense data.
When a backup application sees that the CLN bit in the REQUEST SENSE data
is set, it can prompt the user to clean the drive.
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.
3 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.
3 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.
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
Designing Backup Applications
15
Designing Backup
Applications
Cleaning Tape Heads
DC6000 QIC products, that require the application to estimate what is going
on.
See “Tape Capacity Log Page” under the LOG SENSE command in Chapter 3
of The SCSI Interface, Volume 3 of the HP 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 61.
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 62 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.
16
Designing Backup Applications
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.
Online Help
Good backup applications should include online help that will allow users to
search on key words and to look up error codes reported by the software or
the hardware.
3 Use large SCSI read/write transfer sizes (64 kilobytes is recommended).
3 Incorporate data compression control and report the compression ratios
achieved.
3 Consider where to store directory information depending on the nature of
the application.
3 Only use immediate WRITE FILEMARK commands, but avoid the using
other commands in immediate mode.
3 Use Tape Log information to measure tape quality before backing up
starts.
3 Use the TapeAlert log to prompt the user to take remedial action to avoid
problems.
3 Use “cleaning required” indicators in the software to prompt the user to
clean the drive heads.
3 Allow users to set custom cleaning schedules.
3 Use log files to store Inquiry and Sense Key/Error Code information about
error conditions.
3 Allow users to access drive firmware revision and HBA characteristic
information
3 Include the capability to download firmware.
3 Incorporate simple diagnostic capabilities, such as Write/Read tests and
SCSI bus scanning.
3 Incorporate online help.
Designing Backup Applications
17
Designing Backup
Applications
Design Goals for LTO Backup Applications
18
Designing Backup Applications
Configuration and
Initialization
2
This section covers the following topics:
3 Operating System drivers
3 Inquiry string recovery, finding information about the drive through the
Inquiry command
3 Additional LUN support, for operation with an autochanger device
3 Fibre Channel support
Operating System Drivers
Windows NT
HP have a proprietary, performance-optimized driver for Windows NT 4,
Windows 2000 and Win.NET. It is intended that the driver will be freely
licenced 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
UNIX
Using the NWPA model of NetWare 4.2/5.0, HP are working with Novell to
provide in-box support for HP’s Ultrium products.
See the UNIX 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
Configuration and Initialization
19
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 ”. The 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 be bigger and faster.
Standard INQUIRY Page Data
full-height SCSI
full-height FC
Vendor ID (bytes 8–15)
“HP ”
“HP ”
Product ID (bytes 16–23)
“Ultrium ”
“Ultrium ”
Product ID (bytes 24–31)
“2-SCSI ”
“2-SCSI *”
CYMV
CYMV
Product Revision Level (bytes 32–35)
*This is not a typo!
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:
2nd byte
3rd–6th
bytes
20
Configuration and Initialization
Generation identifier:
“2 ”
Generation 2 (400 GB at 2:1 compression)
“– ”
Hyphen separator (ASCII 2Dh)
“SCSI ” Parallel SCSI interface (and Generation 2 FC drives)
“FC ” Fibre Channel interface, although not for Generation 2
drives
Product Revision Level
1st byte:
Product codename ID:
“F ”
2nd byte
Generation 2 full-height drive
Firmware year:
“0 ”—“9 ” 2000—2009
3rd byte
Firmware month:
“1 ”—“B ” January—December
4th byte
Firmware variant:
“D ”
Standard distribution 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 Mode Sense and Log 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:
To determine the Ultrium format generation:
Use one of the following two methods, of which the second is preferred:
3 Examine the single character following the text “Ultrium ”. A “1 ”
indicates format generation 1 (200 GB capacity at 2:1 compression) and
so on.
3 Preferred method: Use the SCSI Report Density Support command.
For a generation 1 product, the following will be returned:
Primary Density Code:
42h
Vendor-unique Density Code
Assigning Organization: LTO-CVE Linear Tape Open Compliance and
Verification Entity
Density Name:
U28
Ultrium Generation 2, 8-channel
Configuration and Initialization
21
Configuration and
Initialization
Examine only the first eight bytes of the Product ID field (the text “Ultrium ”).
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 Control Interface (ACI)”. The normal tape drive LUN will
always be LUN0, and the automation LUN will always be LUN1.
No other LUNs are available on the drive, although HP is looking to provide
new functionality throught 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 ACI.
22
Configuration and Initialization
Use of Tapes
3
The HP Ultrium user documentation and “Cartridges”, Chapter 9 of the
Hardware Integration Guide, Volume 1 of the HP Ultrium Technical
Refernence Manual, also contain information on Ultrium cartridges.
Timing considerations are discussed in “Time-Out Values” on page 31.
LTO Cartridge Memory
Cartridge Memory has been added to the LTO cartridge for the following
reasons:
3 It speeds up load and unload times by removing the need to read system
areas.
3 It speeds up movement around tape by storing the tape directory (physical
to logical mapping).
3 It increases tape reliability because fewer tape passes are needed.
3 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 some other 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
Ultrium Technical Refernence Manual.
Tape Status and Capacity
Following autoload or a load command, the software can determine the state
of the tape and its capacity from the Tape Usage Log and Tape Capacity Log
Use of Tapes
23
pages retrieved through the LOG 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:
3 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.
3 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:
Units
General
3 Capacities are given in kilobytes of user data.
3 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.
3 System log area, vendor group and EOD areas are not included in
capacities specified. In other words, values are conservative.
3 An allowance for read-after-write retries is made.
Maximum Capacity
3 Maximum capacity values are only valid when the tape has completed a
load sequence. If an immediate mode LOAD is made, LOAD SENSE will not
return valid information until the tape has been successfully loaded and
tape motion has ceased.
Remaining Capacity
24
Use of Tapes
3 The remaining capacity value is the amount of tape remaining calculated
from the current position. As a result, unless the tape is positioned at EOD,
the calculated value ignores the fact that there may be more data written
further up the tape.
3 Remaining capacity values are only valid after the successful completion
of the following commands in non-immediate mode:
LOAD
LOCATE
MODE SELECT
READ
SPACE
VERIFY
WRITE
WRITE FILEMARKS
REWIND
The values after any subsequent command cannot be relied on unless the
command is a sense type that does not cause any tape motion.
Calculations of capacity are inaccurate in two ways:
3 Random errors can be caused by tolerances in tape length, hub diameter,
and so on. These are described below.
3 Systematic errors are caused by ignoring system areas, and so on. They
ensure that the calculated capacity is actually available to the user. It is
usually possible to write considerably more data than the calculated
capacity.
Tape Length Tolerance—Random Errors
Tape length is subject to significant tolerances in its calculation. These derive
from tape thickness, capstan diameter, supply reel hub diameter and supply
reel velocity measurements. Tolerances vary from BOM to EOM due to the
increasing effect of the supply reel hub diameter tolerance.
The existence of these tolerances can cause apparent anomalies in returned
values:
3 Successive loads of the same tape may return different maximum
capacities.
3 At BOM, remaining capacity may be less than maximum capacity. This is
Use of Tapes
Use of Tapes
because maximum capacity is determined when the tape is loaded.
Remaining capacity is determined when the request is made (from the
most recent tape motion data). As a result, the tolerances for these
measurements may cause a discrepancy between the reported values.
25
Responding to Tape Log Data
Note
Software should use the TapeAlert log in preference to the Tape
logs to detect conditions which require the user or host to take
preventative action. See “Monitoring the Condition of the Drive
and Media” on page 62.
These guidelines indicate how host applications should make use of the data
contained in the Tape logs 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.
RAW Retry Counts
Data is read immediately after being written to tape to establish that it has
been written correctly. Increases in RAW retries can be due to four factors:
3 Deterioration in the media
3 Dirty heads
3 Drive malfunction
3 The operating environment
26
Use of Tapes
Corrective Action
The recommended criteria for corrective action are as follows:
RAW Retries > 5% Total groups 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 RAW 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 Tapes
Use of Tapes
27
28
Use of Tapes
Factors Affecting Performance
4
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.
3 Techniques of optimizing performance:
— Detecting the drive’s speed
— Ensuring the recommended minimum transfer sizes
— Using Cartridge Memory instead of tape headers
— Tuning performance using the Performance Log page
3 Time-out values to help you tune timings in backup applications
3 Log pages. This page gives recommendations on the use of information
from the log pages.
3 Factors affecting performance. This page examines factors relating
separately to the host, the drive and the format, and how they can affect
performance.
Ways of Optimizing Performance
HP’s Ultrium drives are high-performance products. The drive’s native speed
is 30 MB/s. With a typical compression ratio of 2:1, this doubles to 60 MB/s.
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.
Factors Affecting Performance
29
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 LOG SENSE command in Chapter 4, “Commands”, of the SCSI
Interface, Volume 3 of the HP Ultrium Technical Reference Manual.
In the Performance Log page (34h), parameter 04h (Native drive speed) gives
the native speed of the drive in units of 100 KB/s. Drives give the value 12Ch,
indicating 30 MB/s.
Ensuring the Recommended Minimum Transfer Sizes
In general, the larger the SCSI transfer sizes for Reads and Writes the better.
When operating in fixed block mode, the actual block size used is largely
immaterial, but the size of each SCSI transfer needs to be maximized. HP
recommends a minimum transfer size of 64 kilobytes in order to minimize
bandwidth lost to SCSI commands overheads. Note that most SCSI cards
working under Windows NT do not support transfer sizes greater than 64 KB
without making changes to the Registry.
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. The
Medium Type field in the Mode Parameter Header is not used and will always
be a blank field.
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 37)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.
30
Factors Affecting Performance
Tuning Performance Using the Performance Log Page
Time-Out Values
SCSI Command
Recommended Time-Out Value
Load
10 minutes
Unload
10 minutes
Rewind (full tape length)
10 minutes
Space/Locate
20 minutes
Erase (long)
5 hours
Erase (short)
5 minutes
Write/Write Filemarks
5 minutes
Read
5 minutes
Read/Write Attribute (MAM), with
1 KB of attribute data
1 minutes
Non-tape movement (such as TEST
1 minutes
UNIT READY, INQUIRY)
Notes:
3 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.
3 All of these values may be subject to change.
3 There is no retension facility.
3 Where relevant, the figures above apply to 200 GB (at 2:1 compression)
media.
Factors Affecting Performance
31
Factors Affecting
Performance
The Performance Log page (34h) contains data that should allow application
software to tune the data-rate being sent to the drive dynamically.
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 Read Attributes command using the Media
Auxiliary Memory (MAM) access 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.
For full details of the Ultrium log pages, see the LOG SENSE command in
Chapter 4, “Commands”, of the SCSI Interface, Volume 3 of the HP Ultiorum
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
Host SCSI performance
Detail
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:
3 Intra-command bus-free time
3 Arbitration and selection
3 Message out
3 Command
Recommendation: To achieve the highest performance, the host must execute the phases that it
controls quickly.
32
Factors Affecting Performance
Performance Factor
Detail
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.
Example: Consider an 8 kilobyte 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.
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.
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.
UNIX seems particularly bad at this.
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.
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. Even on the fast systems at present, the
maximum effective EtherNet bandwidth is only about 1 MB/s, and not all
this bandwidth is usable. However the introduction of 100 MB/s networks
will make this factor of much less significance.
Recommendation: 3 For large datasets, use a tape drive attached directly to the server in
order to maximize throughput.
3 Minimize the number of clients that require backup.
Factors Affecting Performance
33
Factors Affecting
Performance
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.
Host Burst Rate
Performance Factor
Detail
Application Throughput
Some applications move data much quicker than others. Select your
application with care.
Write Commands
Do not interleave write commands with other commands, such as READ
POSITION and LOG SENSE. Do not, for example, attempt to read the
TapeAlert log page during a long write.
Drive-Related Factors
Performance Factor
Drive’s SCSI Performance
Detail
In order to minimize SCSI bus loading, the drive must execute its SCSI
phases quickly. The phases are as follows:
3 Selection
3 Message-out identification
3 Receipt of the command
3 Disconnection
3 Mid-command bus-free time
3 Arbitration and reselection
3 Message-in identification
Recommendation: The host must always ensure the following:
3 Disconnects are enabled
3 Synchronous negotiation is enabled and established between the drive
and the HBA
3 The drive is in buffered mode
3 When reading and verifying, always use the same block size as that in
which the tape is written, otherwise performance will be very seriously
affected.
Transfer Mode
The transfer mode can be fixed or variable, selectable through the MODE
SELECT command.
Fixed Mode: The transfer size is equal to the (block) size multiplied by the
number of records (blocks) in the transfer.
Recommendation: A good transfer size to aim at is 64 kilobytes. For an application that uses
512-byte records, each fixed-mode transfer should transfer 128 records.
Variable Mode: Only one block is transferred at a time. The size of the
block determines the size of the transfer.
34
Factors Affecting Performance
Performance Factor
Detail
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 64-kilobyte transfers.
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 64-kilobyte transfers as a minimum.
Data Compression Ratio
The bandwidth of the data compression engine will determine the drive’s
streaming capabilities based on the compression ratio of the data it is
handling. This is specified in the following table:
Product
Generation 2 full-height drive
Compression Engine
bandwidth
120 MB/s
Max Streaming
Compression Ratio
3:1
The drive will match the throughput of any host up to the maximum native
transfer rate multiplied by the current compression ratio. There will not be
any performance penalty for hosts that are slower than the maximum. This
capability is accomplished with a 64 MB buffer and use of an Adaptive
Tape Speed (ATS) algorithm.
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
frames.
Recommendation: The drive is less likely to stream-read small transfer sizes than it will when
writing transfers of the same sizes.
Raw Tape Data Rate
This is the maximum rate at which data can physically be written to tape.
This is 30 MB/s.
Factors Affecting Performance
35
Factors Affecting
Performance
Recommendation: In variable block mode, the application should use 64-kilobyte blocks.
Performance Factor
Detail
Recommendation: Select the maximum data transfer rate for the type of application used.
Format-Related Factors
Performance Factor
Tapemarks
Detail
Tape marks (filemarks) have many different uses to give a logical 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.
36
Factors Affecting Performance
Supporting Ultrium Features
5
This section covers the following features of HP Ultrium drives:
3 LTO Cartridge Memory
3 Automation Control Interface (ACI)
3 Controlling data compression
3 Other Mode page information: the Buffer Size at EW-EOM field and
partition size
3 Use in tape libraries
Cartridge Memory (LTO-CM)
LTO Contactless 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
3 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.
3 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
Ultrium Technical Reference Manual.
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
Supporting Ultrium Features
37
Needed bit indicates deterioration in the write or read margin of the drive
and hence it is recommended that a cleaning cartridge be 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 be loaded into the drive
immediately. Following a successful clean, the Cleaning Required bit will be
cleared.
Cleaning cartridges can be used for 50 cleanings.
Resetting Drives
The tape drive can be reset by the automation controller via the ACI Reset
command or 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 Ultrium Technical Reference Manual). Two levels of reset via the ACI
interface are provided, namely ACI Reset and Drive Reset. The former resets
the ACI 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 5) can be queued behind outstanding SCSI commands
giving the impression that the drive has stopped responding over the ACI bus.
(All command packages will be still be ack’ed even though the command will
be queued.)
Following an ACI Reset command with reset control set to Drive Reset or after
pulling the ACI_RST_L line low, the drive will behave as if it has powered up
and will go off bus and lose all ACI configurations.
A SCSI interface reset will not affect the ACI interface.
Note that following an upgrade of the drive firmware either via 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.
38
Supporting Ultrium Features
Backup Software
You need backup application software that supports your Ultrium drive and
tape library. For the latest list of backup packages that support your Ultrium
drives, please contact your tape library supplier.
Suitable backup applications will 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, so check the table below for the appropriate reference.
Generation 2 FC drive “HP Ultrium 2-SCSI” (“SCSI” is not a typo!)
Generation 2 SCSI drive “HP Ultrium 2-SCSI”
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 silently and with great agility between the
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.
Support of Data Compression
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. It is not recommended that the host attempts to modify either the
Data Compression Enable (DCE) or Data Decompression Enable (DDE) bits.
Similarly, the Select Data Compression Algorithm (SDCA) field in the Device
Configuration Page (page 10h) is set to 1 (on) and it is not recommended that
this is changed by application software.
Supporting Ultrium Features
39
Supporting Ultrium Features
Drive Model ID String
Other Mode Page Information
Buffer Size at EW-EOM
The Buffer Size at EW-EOM field in the Device Configuration Mode page
(10h) is set to zero as with HP DDS tape drives and other tape technologies.
HP Ultrium tape drives automatically allow sufficient space between EW-EOM
and “physical” EOT to satisfy backup applications.
Partition Size
The LTO Ultrium format only supports a single partition. There is no Medium
Partition Mode page (11h). In the Device Configuration Mode page (10h), the
Change Active Partition (CAP) flag and the Active Partition field should both
be zero since multiple partitions are not supported.
40
Supporting Ultrium Features
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 45.
Code
0h
Sense Key
NO SENSE
Interpretation
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 46.
Action: The host will know what to do with this information, depending upon the
I/O operation at the time. If 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.
The use of Additional Sense 8000h (Cleaning Request) is recommended.
Console Message:
Sense Keys and Codes
41
Code
1h
Sense Key
Interpretation
RECOVERED 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 CHECK CONDITION
having occurred.
For additional sense codes, see “1h—RECOVERED ERROR” on page 46.
Console Message: None
2h
NOT READY
This Sense Key generally means the host will have to wait for the drive to
become READY. Media access is not possible. Click here for additional
sense codes. Also see the Media Access pre-execution check.
Action: The host needs to take one of the following actions:
3 Wait until the drive becomes available.
3 Issue some type of initializing command.
3 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
3h
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 49.
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.
42
Sense Keys and Codes
Code
Sense Key
Interpretation
Console Message: On write (Additional Sense = 0C00h): Write to tape failure
On read (Additional Sense = 70NNh): Decompression exception
On read (otherwise):
4h
HW ERROR
Read from tape failure
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 53.
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
5h
ILLEGAL REQUEST
For additional sense codes, see “5h—ILLEGAL REQUEST” on page 54.
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
6h
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 56.
Also see the Unit Attention pre-execution check.
Sense Keys and Codes
43
Sense Keys and Codes
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.
Code
Sense Key
Interpretation
Action: Recovery depends on the device class and the additional sense code. For
most codes, the software should re-initialize the drive.
The driver should repeat the command. If additional sense code = 290Xh,
the driver should renegotiate SDTR and reconfigure the drive through
MODE SELECT.
Console Message: Tape drive operating conditions may have changed
7h DATA PROTECTION
This is an error if the I/O operation is attempting to access the media in
some manner and failing because the media is write-protected.
For additional sense codes, see “7h—DATA PROTECTION” on page 58.
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
8h
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 58.
Console Message: End-of-Data encountered
Bh
ABORTED
COMMAND
The drive has “given up” on a command. This is normally due to a
problem related to the SCSI bus. 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 59.
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
44
Sense Keys and Codes
Code
Dh
Sense Key
VOLUME
OVERFLOW
Interpretation
Data could not be written because of a lack of remaining space on the
tape. See the Write and Write Filemark commands. Recovery 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 60.
Console Message: Physical End-of-Tape reached, unable to fit
remaining information on tape
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.
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 NOT READY with additional sense of
0402h (LUN not ready, initializing command required). No other
combination of sense key and additional sense may be used to report that
particular condition.
Sense Keys and Codes
45
Sense Keys and Codes
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).
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 info
The drive has no additional sense information for the host. The
flags in the sense data indicate the reason for failure.
Action: see above
00 01 Filemark detected
This indicates one of the following:
3 A READ or SPACE command was terminated early because a
filemark was encountered.
3 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
00 02 End of tape detected
A command completed early because End of Tape or the physical
end of the tape was encountered.
The EOM flag in the sense data will be set.
Action: see above
82 82 Drive requires cleaning
The drive has detected that the heads need to be cleaned to
maintain good operation.
Action: Optionally, log the occurrence for information. It will not be
considered an error and the software will continue.
1h—RECOVERED ERROR
Note
Reporting of recovered errors defaults to OFF.
Action: In all cases, action depends on the device class and operating system.
46
Sense Keys and Codes
Code
Meaning
Comments
37 00 Rounded parameter
The drive needs to round off the value of a parameter sent by
MODE SELECT because it cannot store it to the degree of accuracy
sent by the command.
5D 00 Failure prediction
threshold
Failure Prediction thresholds have been exceeded indicating that a
failure may soon occur.
2h—NOT READY
Code
Meaning
Comments
04 00 LUN Not Ready, no cause This is set if a LOAD/UNLOAD command is received while a load or
to report
unload is already occurring with immediate report on, or initiated
through the front panel. 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.
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.
04 01 LUN in process of
becoming ready
A media access command has been received while a load or
unload is already occurring with immediate report on, or initiated
through the front panel.
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. At this point the command can be
executed.
2 Report to the console that the drive is in the process of loading
tape.
Sense Keys and Codes
47
Sense Keys and Codes
2 Poll waiting until the Additional Sense changes to 3A 00h or
04 02h.
Code
Meaning
Comments
04 02 LUN not ready,
initializing command
required
A cartridge is present in the drive but is not logically loaded.
Action: 1 The software must issue a 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 10 Media Auxiliary Memory is
not accessible
The Media Auxiliary Memory is not accessible
30 03 Cleaning cartridge
A medium-access command has been sent to the drive while a
cleaning cartridge was loaded.
installed.
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.
37 00 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.
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.
48
Sense Keys and Codes
Code
Meaning
Comments
3A 04 Media Auxiliary Memory not MAM is not ready to be accessed. This is likely to be because there is no
ready to access
cartridge present.
3E 00 LUN has not self-
This is set during power-up when it is not possible to send mediumaccess 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.
configured yet
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.
3h—MEDIUM ERROR
Code
Meaning
Comments
00 02 End of tape detected
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.
00 04 Beginning of tape detected
A SPACE command terminated early because it reached BOT.
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 BOP has
been encountered. Also display this information as a console
message.
2 Send any residue information to the calling application.
Sense Keys and Codes
49
Sense Keys and Codes
A 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.
Code
Meaning
Comments
0C 00 Write error
The drive has failed to write data or filemarks to tape. This is
probably due to bad media, but may be hardware-related.
Separate additional sense is reported in the case of an Erase
failure. 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.
0C 0B Media Auxiliary Memory write An error has occurred while attempting to write to MAM. The cartridge
error
should not be used for further backups but should be able to be used for
restoring data.
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.
50
Sense Keys and Codes
Code
Meaning
Comments
11 12 Media Auxiliary Memory read An error has occurred while attempting to write to MAM. The cartridge
error
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: ?
30 01 Cannot read media,
A LOCATE or SPACE command has found the tape is in a format
not supported by the drive.
unknown format
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,
A READ command could not be completed because the logical
format is not correct.
incompatible format
Action: 1 Terminate the current I/O, and return the appropriate error.
3 Prompt the user to eject the cartridge and insert a valid one.
4 Log the incident in the system log.
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.
Sense Keys and Codes
51
Sense Keys and Codes
2 Send a message to the console indicating that the tape is
wrongly formatted.
Code
Meaning
Comments
3B 00 Sequential positioning error The drive has failed to read groups off tape. There are two
possibilities:
3 The current command (such as READ, SPACE, REWIND, or
WRITE) failed to complete successfully.
3 The logical position has been lost. The tape has been
positioned on the other side of the bad groups. The host must
determine if it is worth continuing, which presents a danger
of missing EOD and running into old data. The host must also
decide if the data is old or current.
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 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
Sense Keys and Codes
Code
Meaning
Comments
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.
53 00 Media load/eject failed
A load or eject has 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
media.
3 Advise the user to discard this cartridge and select a new
one.
4 Log the incident in the system log.
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 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 Codes
53
Sense Keys and Codes
4h—HW ERROR
Code
Meaning
Comments
14 03 Manual intervention
A tape is present in the drive but cannot be loaded or unloaded
without manual intervention.
required
Action: ?
40 XX Diagnostic failure on
component XX
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.
Action: see above
44 00 Internal target failure
This code is used to report hardware and firmware related hard
errors that occur when the drive encounters an “impossible”
situation.
Action: see above
51 00 Erase failure
An ERASE command has failed to erase the specified area.
Action: see above
53 00 Media load/eject failed
A load or an eject has failed because of a problem with the tape
cartridge.
Action: see above
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 REQUEST SENSE data bytes and check bytes 15 through 17 to
identify the location of the offending bits or bytes.
4 Refer to the table of Drive Error Codes in Volume 3, The 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 ”.
54
Sense Keys and Codes
Code
Meaning
1A 00 Parameter list length error
Comments
A MODE SELECT parameter list sent to the drive contains one of
the following:
3 An incomplete Mode Parameter header (must be 4 bytes)
3 An incomplete Mode Block Descriptor (must be 0 or 8 bytes)
3 An incomplete Mode page
20 00 Invalid command opcode
The drive does not recognize the opcode of the command it has
received.
24 00 Invalid field in CDB
The drive has detected an invalid field in a command descriptor
block.
25 00 LUN not supported
The command was addressed to a logical unit number that does
not exist.
26 00 Invalid field in Parameter List The drive detected an invalid field among the command
parameters sent during the data phase.
26 01 Parameter not supported
A request for an invalid page number has been sent.
53 02 Medium removal prevented An UNLOAD command failed to eject the tape because medium
removal has been prevented.
The data transferred to the drive during a firmware upgrade is
corrupt or incompatible with the drive hardware.
Sense Keys and Codes
Sense Keys and Codes
82 83 Bad microcode detected
55
6h—UNIT ATTENTION
Code
Meaning
Comments
28 00 Not ready to ready
A tape has been loaded successfully into the drive and is now
ready to be accessed.
transition
Action: 1 The host should be polling, receiving a CHECK 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 ”.
29 00 Power-on, reset or bus
device reset occurred
The drive has powered on or received a SCSI Reset signal or a
Bus Device Reset message since the initiator last accessed it.
Action: The action of the calling application depends on the current I/O
and what the operating system is expecting.
1 As a minimum, the host should re-negotiate for synchronous
negotiation, and reconfigure the drive with any host-specific
operating parameters (burst size, bus activity limit, fixed or
variable mode, and so on).
2 The host should then report to the console that the drive has
been reset.
3 The I/O can be retried if desired.
2A 01 Mode parameters changed
56
Sense Keys and Codes
The Mode parameters for the drive have been changed by a host
other than the one issuing the command. UNIT ATTENTION is set
for all hosts following a MODE SELECT command, apart from the
host that issued the command. This code will only be returned in
a multi-host environment.
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.
3F 01 Microcode downloaded
The firmware in the drive has just been changed by a WRITE
BUFFER command.
3F 11 Media Auxiliary Memory
MAM is accessible but the cartridge is in the Hold position.
accessible
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.
5D FF False informational
exception condition
generated
A TapeAlert 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 CHECK
CONDITION and that the exception will be reported to the
TapeAlert log, no action is necessary.
Sense Keys and Codes
57
Sense Keys and Codes
55 06 Media Auxiliary Memory full
7h—DATA PROTECTION
Code
Meaning
Comments
27 00 Write-protect
This is set if a write operation (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.
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.
8h—BLANK CHECK
Code
Meaning
Comments
00 05 End of Data (EOD) detected A READ or SPACE command terminated early because it
encountered EOD.
Action: 1 Terminate the current I/O and return the appropriate error to
the calling application indicating that EOD has been
encountered.
2 Send a console message saying that EOD has been
encountered.
3 Recovery depends on the calling application and what was
expected.
58
Sense Keys and Codes
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. The most likely cause is a tape with
corrupt format (perhaps from a powerfail when the tape was
being written).
Action: 1 Terminate the current 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 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.
Bh—ABORTED COMMAND
The following action applies to all codes in this group except 00 00h and
44 00h:
Action: 1 Terminate the current I/O and return the appropriate error.
2 Log the incident in the system log.
Code
Meaning
Comments
3D 00 Invalid bits in IDENTIFY
The drive has received an illegal IDENTIFY message at the start of
a command.
43 00 Message error
The number of parity errors or problems on inbound or outbound
messages has exceeded the limit.
message
Note that it is likely that a Bus Release will occur if excessive
parity errors are detected, since the drive will fail to complete the
Status phase and Command Complete message.
45 00 Select or reselect failure
The drive has failed to reselect a host to complete an operation.
CHECK CONDITION state is assumed to exist, even though it was
never reported.
47 00 SCSI parity error
The drive has detected a parity error in an unexpected SCSI state.
It should never do this, so this error should be treated as a
firmware error.
Sense Keys and Codes
59
Sense Keys and Codes
3 Send the console message “SCSI command aborted - low-level
failure on SCSI bus ”.
Code
Meaning
Comments
48 00 Initiator Detected Error
An Initiator Detected Error message has been received, and the
previous phase (the phase in which ATN was asserted) was
invalid.
49 00 Invalid message error
This is set for a number of reasons related to the host and drive
not recognizing each others’ messages.
4A 00 Command phase error
A command could not be executed because too many parity
errors occurred during the Command phase.
4B 00 Data phase error
A command could not be executed because too many parity
errors occurred during the Data-In and Data-Out phases.
4E 00 Overlapped commands
A host has selected the drive even though it already has a
command outstanding.
message received
attempted
Dh—VOLUME OVERFLOW
Code
Meaning
00 02 End of Tape detected
Note
Comments
A WRITE or WRITE FILEMARKS command has encountered EOT or
the physical end of tape. The EOM flag will be set.
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 = NOT READY with
Additional/Qualifying Sense Keys = 04 02. No other
combination of Additional/Qualifying Sense Keys may be used
to report that particular condition.
60
Sense Keys and Codes
7
Exception Handling
These pages cover methods of dealing with certain error conditions and
exceptional circumstances. They include the following:
3 A suggested Escalation Procedure to follow when exceptions occur
3 How to make the most of the TapeAlert facility, see “Supporting
TapeAlert” on page 63
3 How to respond to the ‘Clean’ LED (lit on the front panel when there are
an excessive number of retries or error corrections)
3 The need for a pass-through mode, which can return information or
commands specific to a vendor’s product
3 How drives recover from read and write errors
3 How to use the drive’s ability to read through media errors, so that as
much data can be retrieved from a badly damaged tape as possible
Typical Escalation Procedure
For exception handling, there needs to be a well defined escalation path,
through which the calling application, user, operator or System Supervisor
may take increasingly drastic action to clear any product-related faults.
An escalation procedure is important to allow local recovery where possible,
and to avoid the unnecessary replacement of peripheral devices.
A typical escalation procedure is as follows:
1 Retrieve fault information.
Specific fault information is retrieved from the following sources:
— INQUIRY data, such as firmware revisions
— REQUEST SENSE data, such as Additional Sense Codes and Drive Error
Codes
Exception Handling
61
— MODE SENSE data, such as data on the current configuration
— LOG SENSE data, stored in the drive’s logs
2 Inform the user.
The system gives the user helpful advice by attempting to decode the
returned information, and also allows the user access to the raw data.
3 Allow the user to try recovery.
For any fault, the system allows the user to use simple recovery commands
such as REWIND or LOAD/UNLOAD.
4 Allow the user to reset devices.
If these actions fail, including repeated attempts to retry the operation, the
user should be able to reset devices on the SCSI bus selectively, through
the use of a BUS DEVICE RESET message. This message is specific only to
the device to which it is addressed and will not disturb other devices on the
SCSI bus.
5 Allow the System Supervisor or support person to perform a controlled
hard SCSI Bus Reset.
If the BUS DEVICE RESET message 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 not require a complete shutdown and reboot of the
system.
Monitoring the Condition of the Drive and Media
HP Ultrium drives support TapeAlert (version 3.0), in which the 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’s Guide.
62
Exception Handling
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 is the TapeAlert standard is already fully supported.
As with DDS drives, the benefits of the TapeAlert system can only be fully
realized if the host software supports full SNMP TapeAlert rather than just
Basic TapeAlert. The sending of SNMP traps in response to tape driveinitiated TapeAlert flags allows network management consoles to be aware of
tape drive and backup status.
Monitoring the Condition of the Drive and Media
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.
See the Chapter 4 of Volume 3, The SCSI Interface of the HP Ultrium
Technical Reference Guide for details of the TapeAlert log page.
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 and mandatory for drives that
support cleaning cartridges, as do Ultrium drives.
The flags are grouped into the following sections:
3 Flags 1 to 19: For tape drive write/read management
3 Flags 20 to 25: For cleaning management
3 Flags 26 to 39: For tape drive hardware errors
3 Flags 40 to 49: For tape autoloader errors
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.
Exception Handling
63
Exception Handling
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:
No.
Flag
Type
Recommended Host Message
Cause
Flags for Tape Drive Write/Read Management
The drive is having severe
trouble reading.
1
Read warning
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.
2
Write warning
The drive is having severe
W The tape drive is having problems
writing data. No data has been lost, but trouble writing.
there has been a reduction in the
capacity of the tape.
3
Hard error
W The operation has stopped because an
error has occurred while reading or
writing data which the drive cannot
correct.
4
Media
C
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.
Media performance is severely
1 Copy any data you require from this degraded or the tape can no
longer be written or read.
tape.
Your data is at risk:
This flag is set for any
unrecoverable read/write/
3 Restart the operation with a different positioning error caused by
tape.
faulty media. It is cleared
internally when the media is
ejected.
2 Do not use this tape again.
5
Read failure
C
The tape is damaged or the drive is
The drive can no longer read
faulty. Call the tape supplier’s helpline. 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.
64
Exception Handling
No.
6
Flag
Write failure
Type
C
Recommended Host Message
The tape is from a faulty batch or the
tape drive is faulty:
The drive can no longer write
data to the tape.
1 Use a good tape to test the drive.
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.
2 If the problem persists, call the tape
drive supplier’s helpline.
7
Media life
Cause
W The tape cartridge has reached the end The media has exceeded its
specified life.
of its calculated useful life:
1 Copy any data you need to another
tape.
2 Discard the old tape.
8
9
(not relevant to
Ultrium drives)
A DDS drive has not been able
W The tape cartridge is not data-grade.
Any data you write to the tape is at risk. to read the MRS stripes that
indicate data-grade tape.
Replace the cartridge with a data-grade
tape.
Write-protect
C
Not data grade
You are trying to write to a writeprotected cartridge.
Remove the write-protection or use
another tape.
10 No removal
I
You cannot eject the cartridge because
the tape drive is in use.
A manual or software unload
was attempted when Prevent
Medium Removal was in force.
11 Cleaning media
I
The tape in the drive is a cleaning
cartridge.
A cleaning cartridge is loaded
in the drive.
12 Unsupported
format
I
You have tried to load a cartridge of a Attempted load of an
type that is not supported by this drive. unsupported tape format.
Exception Handling
65
Exception Handling
Wait until the operation is complete
before ejecting the cartridge.
A write command was
attempted to a write-protected
tape.
No.
Flag
13 Recoverable
mechanical
cartridge failure
Type
C
Recommended Host Message
The operation has failed because the
tape in the drive has experienced a
mechanical failure:
1 Discard the old tape.
Cause
The tape has snapped or
suffered some other mechanical
failure in the drive, but the tape
can still be ejected.
2 Restart the operation with a different
tape.
14 Unrecoverable
mechanical
cartridge failure
C
The operation has failed because the
tape in the drive has experienced a
mechanical failure:
1 Do not attempt to extract the tape
cartidge.
The tape has snapped or
suffered some other mechanical
failure in the drive and the tape
cannot be ejected.
2 Call the tape drive supplier’s
helpline.
The LTO-CM chip has failed in
15 Memory chip in W The memory in the tape cartridge has
cartridge failure
failed, which reduces performance. Do cartridge.
not use the cartridge for further write
operations.
A manual or forced eject
occurred while the drive was
writing or reading.
C
The operation has failed because the
tape cartridge was manually demounted while the tape drive was
actively writing or reading.
17 Read-only
format
C
You have loaded a cartridge of a type A write command has been
that is read-only in this drive. The
attempted to a tape whose
cartridge will appear as write-protected. format is read-only in this drive.
18 Tape directory
corrupted on
load
W 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.
The drive was powered down
with a tape loaded, or a
permanent error prevented the
tape directory being updated.
Exception Handling
66
Exception Handling
16 Forced eject
No.
Flag
19 Nearing media
life
Type
I
Recommended Host Message
Cause
The tape cartridge is nearing the end of The tape may have exceeded its
its calculated life. It is recommended that specified number of passes.
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.
Flags for Cleaning Management
20 Clean now
C
The tape drive needs cleaning:
3 If the operation has stopped, eject
the tape and clean the drive.
3 If the operation has not stopped,
The tape drive has detected that
it needs cleaning. The flag is
cleared internally when the
drive is cleaned successfully.
wait for it to finish and then clean the
drive.
Check the tape drive user’s manual for
cleaning instructions.
21 Clean periodic
W The tape drive is due for routine
cleaning:
The drive is ready for a
periodic cleaning.
1 Wait for the current operation to
finish.
2 Use a cleaning cartridge.
Check the tape drive user’s manual for
cleaning instructions.
C
The last cleaning cartridge used in the
tape drive has worn out:
1 Discard the worn-out cleaning
cartridge.
2 Wait for the current operation to
finish.
3 Use a new cleaning cartridge.
The cleaning tape has expired.
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.
Exception Handling
Exception Handling
22 Expired
cleaning media
67
No.
Flag
23 Invalid cleaning
cartridge
Type
C
Recommended Host Message
The last cleaning cartridge used in the
tape drive was an invalid type:
Cause
An invalid cleaning tape type
was used.
1 Do not use this cleaning cartridge in
this drive.
2 Wait for the current operation to
finish.
3 Use a valid cleaning cartridge.
24 Retension
requested
W The tape drive has requested a retension The drive is having trouble
operation.
reading or writing that will be
resolved by a retension cycle.
25 Dual-port
interface error
W A redundant interface port on the tape
drive has failed.
One of the interface ports in a
dual-port configuration (in
other words, Fibre Channel)
has failed.
Flags for Tape Drive Hardware Errors
W A tape drive cooling fan has failed.
A fan inside the drive
mechanism or enclosure has
failed.
27 Power supply
failure
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.
A redundant PSU has failed
inside the tape drive enclosure
ot rack subsystem.
28 Power
consuption
W The tape drive power consumption is
outside the specified range.
The tape drive power
consumption is outside the
specified range.
29 Drive
maintenance
W Preventive maintenance of the tape
drive is required.
The drive requires preventive
maintenance (not cleaning).
Exception Handling
26 Cooling fan
failure
Check the tape drive user’s manual for
preventive maintenance tasks or call the
tape drive supplier’s helpline.
Exception Handling
68
No.
Flag
30 Hardware A
Type
C
Recommended Host Message
The tape drive has a hardware fault:
1 Eject the tape or magazine.
2 Reset the drive.
Cause
The drive has a hardware fault
from which it can recover
through a reset.
3 Restart the operation.
31 Hardware B
C
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.
32 Interface
W The tape drive has a problem with the
application client interface:
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.
The drive has identified an
interface fault.
1 Check the cables and cable
connections.
2 Restart the operation.
33 Eject media
C
The operation has failed:
Error recovery action.
1 Eject the tape or magazine.
2 Insert the tape or magazine again.
3 Restart the operation.
34 Download fail
W The firmware download has failed
because you have tried to use the
incorrect firmware for this tape drive.
Firmware download failed.
35 Drive humidity
W Environmental conditions inside the tape The drive’s humidity limits have
drive are outside the specified humidity been exceeded.
range.
36 Drive
temperature
W Environmental conditions inside the tape The drive is experiencing a
drive are outside the specified
cooling problem.
temperature range.
Exception Handling
69
Exception Handling
Obtain the correct firmware and try
again.
No.
Flag
Type
Recommended Host Message
Cause
37 Drive voltage
W The voltage supply to the tape drive is
outside the specified range.
Drive voltage limits have been
exceeded.
38 Predictive
failure
C
39 Diagnostics
required
The drive may have a hardware
W The tape drive may have a hardware
fault. Run extended diagnostics to verify fault that may be identified by
extended diagnostics (using a
and diagnose the problem.
SEND DIAGNOSTIC command).
Check the tape drive user’s manual for
instructions on running extended
diagnostic tests.
Failure of the drive’s hardware
A hardware failure of the drive is
predicted. Call the tape drive supplier ‘s is predicted.
helpline.
Flags for Tape Autoloader Errors
Flags 40–49 are not currently used.
50 Lost statistics
W Media statistics have been lost at some
time in the past.
51 Tape directory
invalid at
unload
W The tape directory on the tape cartridge An error has occurred
just unloaded has been corrupted. File preventing the tape directory
search performance will be degraded. being updated on unload.
The tape directory can be rebuilt by
reading all the data.
52 Tape system
area write
failure
C
The tape just unloaded could not write
its system area successfully:
1 Copy the data to another tape
cartridge.
The drive or library has been
powered on with a tape
loaded.
Write errors occurred while
writing the system area on
unload.
2 Discard the old cartridge.
53 Tape system
area read
failure
C
The tape system area could not be read Read errors occurred while
successfully at load time.
reading the system area on
load.
Copy the data to another tape
cartridge.
70
Exception Handling
No.
Flag
54 No start of data
Type
C
Recommended Host Message
Cause
The start of data could not be found on The tape has been damaged,
bulk erased, or is of an
the tape:
incorrect format.
1 Check that you are using the correct
format tape.
2 Discard the tape or return the tape to
your supplier.
55 Loading failure
C
The drive is unable to load the
The operation has failed because the
media cannot be loaded and threaded. cassette and thread the tape.
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.
56 Unrecoverable
load failure
C
The operation has failed because the
tape cannot be unloaded:
The drive is unable to unload
the tape.
1 Do not attempt to extract the tape
cartridge.
2 Call the tape driver supplier’s help
line.
57 Automation
interface failure
C
The tape drive has a problem with the
automation interface:
The drive has identified a fault
in the automation interface.
1 Check the power to the automation
system.
2 Check the cables and cable
connections.
58 Firmware
failure
Exception Handling
3 Call the supplier’s helpline if the
problem persists.
W 4 The tape drive has reset itself due to There is a firmware bug.
a detected firmware fault. If the
problem persists, call the supplier’s
helpline.
Flags 59–64 are not currently used.
Exception Handling
71
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:
3 At power-on.
3 When specified corrective action has been taken, such as using a cleaning
cartridge.
3 When the TapeAlert Log page is read.
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:
3 Use the MODE SELECT TapeAlert Mode page to enable “Check Condition”
mode. This means that the tape drive reports CHECK CONDITION on the
next SCSI command after one or more TapeAlert flags are set. When
CHECK CONDITION is received, the host software should behave as
follows:
— It issues a REQUEST SENSE command. Additional sense of 5D00h
indicates that the CHECK CONDITION was caused by TapeAlert. This
enables the software to distinguish CHECK CONDITIONs caused by
actual errors and those resulting from a TapeAlert flag being set.
— The software then reads the TapeAlert log page to discover which flags
are set (even for CHECK CONDITIONs caused by actual errors).
Note that when CHECK CONDITION results from TapeAlert, the command
that reported the CHECK CONDITION is not in error and will have
72
Exception Handling
completed successfully. It follows that the software should not repeat the
command.
3 Read the TapeAlert log page using LOG SENSE at the following times:
— Immediately after a SCSI CHECK CONDITION/REQUEST SENSE cycle.
— At the end of each tape in cases where a backup or restore spans
multiple tapes. If the tape is ejected, the host must read the TapeAlert
log page before this occurs.
— 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.
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 notification 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 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 Handling
73
Exception Handling
— The software label of the tape cartridge when relevant.
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.
Responding to the ‘Clean’ LED
Note
We recommend 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 RAW
retries, the ‘Clean’ LED is lit. If this happens, a user should follow this
procedure:
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 or by power-cycling.
74
Exception Handling
Providing Pass-Through Mode
It is important for Drivers and Logical Device Managers to provide a passthrough mode that can return information or commands specific to a vendor’s
product. The need for this is two-fold:
3 Systems must support a great variety of new devices.
3 For the most part, all tape drives are similar, and Drivers and Logical
Device Managers tend to provide connectivity based on the assumption
that 80% of all SCSI tape drives behave identically.
A pass-through mode offers the following advantages:
3 It allows the peripheral manufacturer to provide value-added diagnostics
and support applications over and above those that may be shipped with
a system or application.
3 It allows System Supervisors and operators to take advantage of specific
product features that are otherwise excluded because the driver or
manager only caters for 80% of SCSI drives.
3 It allows technical support people to have access to low-level device-
specific information that otherwise is likely to be unavailable.
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 or 10-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.
or
Allocate a read buffer or file for any data associated with the SCSI
command that will be returned by the drive.
4 Launch the command.
5 The driver should use its standard CHECK CONDITION and REQUEST
SENSE routines to report whether the command completed successfully or
not. The caller must have access to the raw REQUEST SENSE data.
6 View any returned data.
Exception Handling
75
Exception Handling
3 Link the command and data buffers.
Recovering from Write and Read Errors
Write Recovery
HP Ultrium drives ensure data integrity by performing Read-After-Write
(RAW) checks while data is written to tape. The drive still maintains
streaming.
Read Recovery
If all else fails, the host can attempt to read through the media error. See the
next section, “Reading Through Media Errors”.
Reading Through Media Errors
In order to cope with exceptional situations, the drive can read through media
errors in an attempt to recover as much data from tape as possible. Media
errors in this case mean hard media errors, from which the drive’s error
correction techniques are unable to recover. The media errors may be caused
by poor storage or tape abuse.
The following scenario shows how the read-through facility could be used:
1 Media errors of this type can be identified in response to a READ
command by a sense key of MEDIUM ERROR, and one of the following
additional sense codes:
11 00h
Unrecovered read error
11 01h
Read retries exhausted
31 00h
Format command failed
09 00h
Track following error
15 02h
Positioning error detected by read of medium
2 When one of these errors occurs, the drive automatically attempts to read
through several groups of data on tape (the exact number depends on the
drive).
76
Exception Handling
3 The host should continue to issue READ commands until no CHECK
CONDITIONs are returned, or EOD is encountered. This means the drive
has reached good data, and the host should concentrate on redefining the
actual logical position.
4 In the extremely unlikely situation that a large number of groups are
affected by damaged media, the host could still continue to issue READ
commands. The drive will then read through further sets of data. The
process is continuous providing that the host issues further READs. In doing
this, much time is involved in recovery, and a time-out value of 7 minutes
per READ is strongly recommended.
5 Once a READ completes successfully, the host should continue to issue
READ commands, or issue a SPACE to filemark command. By reading the
header after the filemark, the host can establish which files are
unrecoverable, but more importantly, can continue to recover data
beyond the damaged area of tape.
Note
It is not recommended to use the SPACE command to move over
a media error.
In addition, if a filemark is involved in the media error, the
SPACE command will not complete successfully. This is because
the filemark count will be inconsistent and the drive will start
searching in the flawed area trying to find the missing filemark.
A filemark can only be detected on a SPACE command if the
whole group containing the filemark can be read.
Exception Handling
Exception Handling
77
78
Exception Handling
glossary
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 MODE SELECT
Parameter List 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.
EOM
End Of Media format. The last usable point on the tape.
EW-EOM
filemark
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.
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.
glossary
79
filemark count
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.
LVD
Low-Voltage Differential. See SCSI.
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.
RAW
read-after-write
SCSI
see read-after-write
RAW improves data integrity by reading data immediately after it is written
and writing the data again if an error is found.
Small Computer System Interface—a standard command specification and
command set that enables computers and peripherals to communicate with
each other. HP’s Ultrium drives adhere to the SCSI specifications (see
Chapter 1, “Interface Implementation” in Volume 3, The SCSI Interface, of the
HP Ultrium Technical Reference Manual) and support all features required by
those standard.
Single-Ended and Low Voltage Differential SCSI
These terms define how the signals are transmitted along the cable.
With single-ended (SE) SCSI, each signal travels over a single wire and each
signal’s value is determined by comparing the signal to a paired ground wire.
Signal quality tends to decrease over longer cable lengths or at increased
signal speed.
80
glossary
Narrow and Wide, Fast, Ultra and Ultra2 SCSI
Narrow SCSI devices can transfer data one byte at-a-time (and are
sometimes called “8-bit SCSI” devices). They can conform to either the SCSI-2
or SCSI-3 protocols. They have a 50-pin connection to the SCSI bus.
Wide SCSI devices can transfer two bytes of data simultaneously (“16-bit
SCSI”). They usually have a single, 68-pin connection to the SCSI bus. (This
physical arrangement is part of the SCSI-3 specification.) They may support
either SCSI-2 or SCSI-3 protocols. Wide and narrow devices can
simultaneously be connected to the same bus without problem, provided
certain rules are followed.
Fast SCSI can transfer data at up to 20 MB/s wide, using a cable of up to
6 meters total length.
Ultra SCSI can transfer data at up to 40 MB/s wide, but the cable length
cannot exceed 3 meters (it is also known as “Fast20”).
Ultra2 SCSI can transfer data at up to 80 MB/s wide, using a cable of up to
25 meters total length for a single device, or up to 12 meters for two or more
devices (it is also known as “Fast40”).
Ultra3 or Ultra160 can transfer data at up to 160 MB/s wide. Cable lengths
are as for Ultra2.
Ultra4 or Ultra320 will transfer at up to 320 MB/s.
Ultra SCSI supports both SE and LVD interfaces. In normal situations, slower
devices can coexist with faster devices, and narrow devices can be used on
the same SCSI bus as wide devices using a suitable adapter.
HP’s Generation 1 Ultrium drives are Ultra2, wide SCSI-3 compatible devices.
They can be used with both LVD and SE host bus adapters.
Generation 2 Ultrium drives are Ultra160, wide SCSI-3 compatible.
single-ended
TapeAlert
see SCSI
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.
glossary
81
glossary
With low voltage differential (LVD) signaling, signals travel along two wires
and the difference in voltage between the wire pairs determines the signal
value. This enables faster data rates and longer cabling with less susceptibility
to noise than SE signaling and reduced power consumption.
82
glossary
index
Index
A
D
ACI 22
additional sense codes 45
ANSI 3, 79
autochangers 22
data compression 79
controlling 13, 14, 39
effect on performance 35
data transfer size 13
decompression 79
design goals 17
Diagnostic logs 16
diagnostics, failure 54
displaying drive information 16
documents, related 3
drive
checking integrity 17
detecting speed 30
displaying information 16
initialization 19
monitoring condition 63
drive technology family 21
drivers 19
B
backup applications 13
block mode
fixed-length 13
variable-length 14
block size, maximum 30
BOM 79
BOT 49
buffer
maximizing use 13
size at EW-EOM 40
buffered mode 79
C
capacity
of tape 23
remaining 24
cartridge memory 23, 37
using instead of headers 30
cartridges, see tapes
Clean LED 74
cleaning 46, 74
cleaning cartridge 48
cleaning tape heads 15
commands, non-immediate 14
compression 79
controlling 39
ratio 79
configuration 19
E
ECMA 79
eject, failed 53, 54
EOD 79
detected 58
not found 59
EOM 46, 49, 79
EOT 60
erase failure 54
errors
hard 80
parity 59
read 50
write 50, 52
escalation procedure 61
Index
83
EW-EOM 79
exception handling 61
F
faults, predicting 16, 62
filemarks 79
count 80
detected 46
use of 36
firmware revision 16
fixed-length block mode 13
flags, TapeAlert 63
format
corrupt 51
incompatible 51
front panel LEDs 74
FRU 80
G
generation, finding 21
H
hard error 80
help 17
host 80
I
initialization 19
INQUIRY command 16
inquiry string recovery 19
L
LEDs, Clean 74
load
count 26
failed 53, 54
loading 80
LOG SENSE data 24
logs
Diagnostic 16
performance 31
supporting 32
Tape 15, 23, 26
Tape Capacity 23
84
Index
TapeAlert 16, 26, 63, 81
LTO-CM 23, 37
using instead of headers 30
LUNs 80
not ready 47
support for additional 22
M
MAM
write error 50, 51
media errors, reading through 76
media, see tapes
memory, cartridge 23
microcode 57
mode
fixed-length block 13
pass-through 75
variable-length block 14
mode parameters 56
monitoring condition 16, 62
monitoring tape use 15
N
NetWare drivers 19
non-immediate commands 14
O
offline 80
online 80
online help 17
operating system drivers 19
optimizing performance 13, 29
P
parity error 59
partition size 40
pass-through mode 75
performance factors 29
drive-related 34
format-related 36
host-related 32
performance log 31
problems, predicting 16, 62
product ID 20
product revision level 20, 21
R
RAW 76, 80
retries 74
retry counts 26
read errors 50
reading through 76
recovery 76
reading the TapeAlert log 73
remaining capacity 24
revision level 20, 21
S
U
Ultrium
finding the format generation 21
supporting features 37
UNIX drivers 19
V
variable-length block mode 14
vendor ID 20
W
Windows NT drivers 19
write errors 50, 52
recovery 76
write-protect 58
T
Tape Capacity log 23
tape heads, cleaning 15
Tape log 15, 23
responding to 26
TapeAlert log 16, 26, 63, 81
reading 73
tapemarks, use of 36
tapes
capacity 23
cleaning cartridge 48
faulty 53
identifying 30
Index
85
Index
SCSI 3, 80
sense keys
ABORTED COMMAND 44, 59
BLANK CHECK 44, 58
DATA PROTECTION 44, 58
HW ERROR 43, 53
ILLEGAL REQUEST 43, 54
MEDIUM ERROR 42, 49
NO SENSE 41, 46
NOT READY 42, 47
RECOVERED ERROR 42, 46
UNIT ATTENTION 43, 56
VOLUME OVERFLOW 45, 60
speed, detecting 30
status of tape 23
support for additional LUNs 22
length tolerance 25
monitoring condition 63
monitoring use 15
status 23
use of 14
using 23
tests, read/write 17
time-out values 31
transfer size 13, 30
effect on performance 35
troubleshooting 16
86
Index
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement