Computer Drive User Manual
Table Of Contents
- Chapter 1. HPSS 7.1 Configuration Overview
- Chapter 2. Security and System Access
- Chapter 3. Using SSM
- 3.1. The SSM System Manager
- 3.2. Quick Startup of hpssgui
- 3.3. Configuration and Startup of hpssgui and hpssadm
- 3.4. Multiple SSM Sessions
- 3.5. SSM Window Conventions
- 3.6. Common Window Elements
- 3.7. Help Menu Overview
- 3.8. Monitor, Operations and Configure Menus Overview
- 3.9. SSM Specific Windows
- 3.10. SSM List Preferences
- Chapter 4. Global & Subsystem Configuration
- 4.1. Global Configuration Window
- 4.2. Storage Subsystems
- 4.2.1. Subsystems List Window
- 4.2.2. Creating a New Storage Subsystem
- 4.2.3. Storage Subsystem Configuration Window
- 4.2.3.1. Create Storage Subsystem Metadata
- 4.2.3.2. Create Storage Subsystem Configuration
- 4.2.3.3. Create Storage Subsystem Servers
- 4.2.3.4. Assign a Gatekeeper if Required
- 4.2.3.5. Assign Storage Resources to the Storage Subsystem
- 4.2.3.6. Create Storage Subsystem Fileset and Junction
- 4.2.3.7. Migration and Purge Policy Overrides
- 4.2.3.8. Storage Class Threshold Overrides
- 4.2.4. Modifying a Storage Subsystem
- 4.2.5. Deleting a Storage Subsystem
- Chapter 5. HPSS Servers
- 5.1. Server List
- 5.1. Server Configuration
- 5.1.1. Common Server Configuration
- 5.1.1. Core Server Specific Configuration
- 5.1.2. Gatekeeper Specific Configuration
- 5.1.3. Location Server Additional Configuration
- 5.1.4. Log Client Specific Configuration
- 5.1.1. Log Daemon Specific Configuration
- 5.1.2. Migration/Purge Server (MPS) Specific Configuration
- 5.1.3. Mover Specific Configuration
- 5.1.3.1. Mover Specific Configuration Window
- 5.1.3.1. Additional Mover Configuration
- 5.1.3.1.1. /etc/services, /etc/inetd.conf, and /etc/xinetd.d
- 5.1.3.1.2. The Mover Encryption Key Files
- 5.1.3.1.3. /var/hpss/etc Files Required for Remote Mover
- 5.1.3.1.1. System Configuration Parameters on IRIX, Solaris, and Linux
- 5.1.3.1.1. Setting Up Remote Movers with mkhpss
- 5.1.3.1.2. Mover Configuration to Support Local File Transfer
- 5.1.1. Physical Volume Repository (PVR) Specific Configuration
- 5.1.1. Deleting a Server Configuration
- 5.1. Monitoring Server Information
- 5.1.1. Basic Server Information
- 5.1.1. Specific Server Information
- 5.1.1.1. Core Server Information Window
- 5.1.1.1. Gatekeeper Information Window
- 5.1.1.1. Location Server Information Window
- 5.1.1.2. Migration/Purge Server Information Window
- 5.1.1.3. Mover Information Window
- 5.1.1.1. Physical Volume Library (PVL) Information Window
- 5.1.1.2. Physical Volume Repository (PVR) Information Windows
- 5.1. Real-Time Monitoring (RTM)
- 5.2. Starting HPSS
- 5.1. Stopping HPSS
- 5.2. Server Repair and Reinitialization
- 5.1. Forcing an SSM Connection
- Chapter 6. Storage Configuration
- 6.1. Storage Classes
- 6.2. Storage Hierarchies
- 6.3. Classes of Service
- 6.4. Migration Policies
- 6.5. Purge Policies
- 6.6. File Families
- Chapter 7. Device and Drive Management
- Chapter 8. Volume and Storage Management
- 8.1. Adding Storage Space
- 8.2. Removing Storage Space
- 8.3. Monitoring Storage Space
- 8.4. Dealing with a Space Shortage
- 8.5. Volume Management
- 8.6. Monitoring and Managing Volume Mounts
- 8.7. New Storage Technology Insertion
- Chapter 9. Logging and Status
- Chapter 10. Filesets and Junctions
- Chapter 11. Files, Directories and Objects by SOID
- Chapter 12. Tape Aggregation
- Chapter 13. User Accounts and Accounting
- Chapter 14. User Interfaces
- Chapter 15. Backup and Recovery
- Chapter 16. Management Tools

• Auto Stage Retry. When this flag is turned on, and a valid secondary copy of the data exists, and a
stage from the primary copy fails, HPSS will automatically retry the stage using the secondary copy.
• Auto Read Retry. When this flag is turned on, and a valid secondary copy of the data exists, and an
attempt to read the first copy fails, HPSS will automatically retry the read using the secondary copy.
This retry will not be attempted if any data was transferred between the client and HPSS during the
failure.
For Auto Stage Retry and Auto Read Retry to work properly, the COS must contain at
least two copies of the files and at least one valid second copy must have been created
during HPSS migration processing. If any tape storage class in the COS has a migration
policy, that policy must be configured for Migrate Files or Migrate Files and Purge. A
storage class whose migration policy is configured for Migrate Volumes or Migrate
Volumes and Whole Files will not be considered as a valid alternate copy from which
to retry a failed read or stage operation. Any storage class to which such a storage class
migrates, directly or indirectly, will not be considered as a valid retry candidate, either.
•
Truncate Final Segment. A flag to influence whether the final segment of files will be truncated at
file close time in order to save disk space.
When a file is closed, its final segment may be truncated to the smallest segment size which is valid
for the storage class and will still hold the actual data written to the segment. This size will be a
power-of-two multiple of the minimum segment size configured for the storage class. The segment
will later be expanded to its original size if the file is re-opened and new data is written to the segment
or to a point in the file beyond the segment. Truncation of the final segment is performed only on the
file data at the top level of the storage hierarchy and only if the storage class at the top level is disk.
Truncation of the final segment is controlled by the NOTRUNC_FINAL_SEG flag in the bitfile
descriptor; if this flag is off, truncation is performed at file close, and if it is on, truncation is not
performed.
The COS Truncate Final Segment flag influences whether the NOTRUNC_FINAL_SEG flag in the
bitfile descriptor is set at file creation time. The user may specify a creation hint to turn truncation off
even if the COS allows truncation, but he may not turn truncation on if the COS prohibits it. If the
COS Truncate Final Segment flag is on, then at file creation time the NOTRUNC_FINAL_SEG flag
will be set to off, unless the user specifies the NOTRUNC_FINAL_SEG hint. If the COS Truncate
Final Segment flag is off, then the NOTRUNC_FINAL_SEG flag will be set to on at file creation
time, and the user may not override this.
Users may also turn on the NOTRUNC_FINAL_SEG flag in their own files after file creation by use
of the hpss_SetFileNotrunc client api function, but they may not turn off the flag. Authorized callers
may use the hpss_SetFileNotrunc function to turn the flag on or off for any file. Authorized callers
are those such as hpssssm who have CONTROL permission on the core server's Client Interface ACL.
Advice – Truncation of the final segment normally saves disk space. However, if a file is frequently
written, closed, re-opened, and appended to, the repeated truncation and re-expansion of the final
segment could result in fragmentation of the disk, as the space for the expansion is not guaranteed to
be adjacent to the original segment. In addition, it is possible that the only extent available for the
expansion is larger than needed, so some space could actually be wasted.
If it is known that the users of a COS do frequent appends to their files, it is better to turn the COS
Truncate Final Segment flag off to avoid these potential fragmentation and space waste issues. If no
HPSS Management Guide November 2009
Release 7.3 (Revision 1.0) 177