HP OpenVMS Version 8.2 New Features and
3.6 Host-Based Adapter (HBA) Support
The following sections describe the HBA support in OpenVMS Version 8.2:
- Fibre Channel Host-Based Adapter
- Ultra SCSI Host-Based Adapter
3.6.1 Fibre Channel HBA Support on OpenVMS I64 and OpenVMS Alpha Systems
OpenVMS I64 Version 8.2 supports only the dual-port 2GB/1GB Fibre
Channel Universal PCI-X HBA (A6826A) for external connection to Fibre
Channel storage. This HBA is supported by a new driver, the PGQDRIVER,
which has been implemented as a port driver to the existing DKDRIVER.
The PCI-X 1-port FCA2404 2GB adapter (AB232A or KGPSA-EA, aka LP9802
for AlphaServers) supported on the OpenVMS I64 V8.1 Evaluation Release
for Integrity servers is no longer supported on OpenVMS V8.2 or later.
For external Fibre Channel storage, customers should plan accordingly
to replace AB232A or KGPSA-EA FC adapters with the A6826A adapter.
OpenVMS Alpha Version 8.2 supports the new Fibre Channel HBA, the
LP10000. The LP10000 is supported on the AlphaServer DS and ES server
families. The LP10000 is available in a single-channel or dual-channel
The LP10000 is supported on earlier versions of OpenVMS Alpha (Version
7.3-1, and Version 7.3-2) by means of patch kits and by firmware
The version-specific patch kits with a root name of FIBRE_SCSI will be
available at the following website:
Under Service tools, select either Patches for OpenVMS or FTP site for
For more detailed information about these HBAs, refer to the hardware
documentation for your HP Integrity server or for your adapter.
3.6.2 Ultra SCSI HBA Support on OpenVMS I64 Systems
OpenVMS I64 V8.2 supports the following Ultra SCSI HBAs on HP Integrity
- Ultra-160 SCSI dual channel (A6829A)---on HP rx4640 Integrity
- Ultra-320 SCSI dual channel (A7173A)---embedded in HP Integrity
rx2600 server and HP Integrity rx1600 server
External SCSI storage requires the addition of an Ultra-160 dual-port
SCSI adapter (A6829A), which can be housed in DS2100, MSA30, or
4200/4300 series shelves.
OpenVMS does not support shared SCSI storage using these adapters in an
OpenVMS Cluster system consisting of either OpenVMS I64 systems only,
or a mix of OpenVMS I64 systems and OpenVMS Alpha systems.
These adapters are also supported by HP on the following operating
systems from HP: HP-UX, Linux, and the Microsoft XP 64-bit operating
For more detailed information, refer to the hardware documentation that
accompanies your HP Integrity server.
3.7 System Analysis Tools Enhancements
The System Analysis Tools have been enhanced with new commands and new
features for use with both Alpha and I64 systems. The new commands and
features make it easier to analyze an OpenVMS system.
For more detailed information, refer to the HP OpenVMS System Analysis Tools Manual.
3.7.1 New and Enhanced SDA Commands
The following commands have been enhanced or are new for I64 use:
- SET CPU
- SET OUTPUT
- VALIDATE TQEIDX
3.7.2 System Service Logging
System Service Logging records information about system service
activity within a process. It is intended for use when troubleshooting
Logging is turned on with the SET PROCESS/SSLOG=(STATE=ON) command.
Logging is stopped with the SET PROCESS/SSLOG=(STATE=UNLOAD) or
(STATE=OFF) command. Logged information is displayed with the
ANALYZE/SSLOG command. SSLOG.DAT is the default file where logged
information is captured.
For additional information, see the HP OpenVMS System Analysis Tools Manual.
3.8 System Parameters
The system parameters in the following sections are new in Version 8.2.
The last section lists the system parameters that have been changed in
Version 8.2. For more detailed information, refer to the HP OpenVMS System Management Utilities Reference Manual.
3.8.1 New System Parameters
The following parameters are new in OpenVMS Version 8.2:
ERLBUFFERPAG_S2 specifies the amount of S2
space memory to allocate for each S2 space error log buffer requested
by the ERRORLOGBUFF_S2 parameter.
ERRORLOGBUFF_S2 specifies the number of S2 space error log buffers
reserved for system error log entries. Each buffer is ERLBUFFERPAG_S2
in length. If ERRORLOGBUFF_S2 is too low, messages might not be written
to the error log file. If it is too high, the buffers can consume
unnecessary physical pages.
The purpose of SCSI_ERROR_POLL is to cause OpenVMS to send a SCSI
Test Unit Ready command every hour to each SCSI disk, in an attempt to
force latched errors to become unlatched and to be reported
immediately. SCSI_ERROR_POLL has a default value of 1. However, it can
be set to 0 by the user in order to stop the error polling activity.
Special parameter reserved for HP use.
- SHADOW_HBMM_RTC (Alpha and I64 only)
SHADOW_HBMM_RTC controls the interval the system waits between the
checking of reset thresholds for shadow sets that have Host-Based
Minimerge (HBMM) bitmaps. If the reset threshold is exceeded, the
bitmap is zeroed.
When a copy or merge operation is needed on a shadow set that is
mounted on many systems, the Shadowing facility attempts to perform the
operation on a system that has a local connection to all the shadow set
members. Shadowing implements the copy or merge operation by adding a
time delay based on the number of shadow set members that are
MSCP-served to the system. No delay is added for local members;
therefore, a system with all locally accessible shadow set members
usually performs the copy or merge before a system on which one or more
members is served---and is therefore delayed---does.
allows the system manager to adjust the delay that Shadowing adds. By
default, the delay is 30 seconds for each MSCP-served shadow set
member. The valid range for the specified delay is 0 through 65,535
When a shadow set is mounted on a system, the value of
SHADOW_PSM_DLY is used as the default shadow set member recovery delay
for that shadow set. To modify SHADOW_PSM_DLY for an existing shadow
set, refer to the SET SHADOW/RECOVERY_OPTIONS=DELAY_PER_SERVED_MEMBER=n
- SHADOW_REC_DLY (Alpha and I64 only)
The value of SHADOW_REC_DLY is added to the value of the
RECNXINTERVAL parameter to determine the length of time a system waits
before it attempts to manage recovery operations on shadow sets that
are mounted on the system, using the priority of the shadow sets.
- SHADOW_SITE_ID (Alpha and I64 only)
SHADOW_SITE_ID allows a system manager to define a site value,
which Volume Shadowing uses to determine the best device to perform
reads, thereby improving performance. The system manager can now define
the site value to be used for all shadow sets mounted on a system.
- SYSSER_LOGGING (Alpha and I64 only)
A value of 1 for SYSSER_LOGGING enables logging of system service
requests for a process. The default is 1. This parameter is dynamic.
- TTY_DEFCHAR3 (Alpha and I64 only)
TTY_DEFCHAR3 allows a user to set a bit so that the OpenVMS
terminal driver remaps Ctrl/H to Delete. HP recommends that you
not set this bit as a systemwide default.
When this bit is set, the OpenVMS terminal console remaps CTRL/H to
- VHPT_SIZE (I64 only)
VHPT_SIZE is the number of kilobytes to allocate for the Virtual
Hash Page Table (VHPT) on each CPU in the system:
- 0 indicates that no VHPT is allocated.
- 1 indicates that OpenVMS chooses a default size that is
appropriate for your system configuration.
3.8.2 Changed System Parameters
Definitions of the following system parameters have been changed in
OpenVMS Version 8.2:
In addition to the parameters in this list, many parameters that were
formerly designated as Alpha only are now Alpha and
For more detailed information, refer to the HP OpenVMS Version 8.2 Release Notes and
HP OpenVMS System Management Utilities Reference Manual.
3.9 Additional Time Zones Added to Database
In OpenVMS Version 8.2, 540 time zones are provided based on the
time-zone public database, named tzdata2003e, placed in
. Existing time zones are updated, and there are 204 new time zones
added to the database. For a list of the new time-zone names, see
HP OpenVMS System Manager's Manual.
3.10 Volume Shadowing for OpenVMS New Features
Volume Shadowing for OpenVMS introduces the following new features in
- Host-Based Minimerge (HBMM)
HBMM is designed to improve shadow
set merge operations. HBMM uses a bitmap to track the blocks that
changed. HBMM compares and reconciles only those sections of the shadow
set where changes occurred since the bitmap was reset.
An HBMM policy keyword, RESET_THRESHOLD, is used to specify the
number of blocks that can be changed before the bitmap is cleared. A
new system parameter, SHADOW_HBMM_RTC, is used to specify how
frequently the bitmap is checked to determine if the number of changed
blocks exceeds the RESET_THRESHOLD setting. If it does, the bitmap is
cleared. HBMM operations can be significantly faster than full merges,
depending on the value of RESET_THRESHOLD.
- Prioritizing merge and copy operations
You can now control the order in which merge and copy operations occur
on the remaining systems in a cluster after a system failure. The new
/PRIORITY=n qualifier to the SET SHADOW command can be used to
assign different priorities to all mounted shadow sets. Assigning
higher priorities to the most important volumes ensures that they are
merged or copied before less important volumes, which are assigned a
You can dynamically increase or decrease the SHADOW_MAX_COPY setting as
circumstances change in the cluster. Without having to dismount the
shadow sets, you can use the new EVALUATE_RESOURCES qualifier to the
SET SHADOW command to cause a system to adjust its workload to a new
SHADOW_MAX_COPY value or to new priorities specified with the SET
SHADOW/PRIORITY=n DSAn command.
A new system parameter, SHADOW_REC_DLY, enables you to predict the
order that each system attempts to manage copy or merge operations
after a system failure. You can assign the shortest delays to the
systems that are best able to perform recovery operations.
- Licensing Volume Shadowing for OpenVMS
You can purchase
capacity licenses, per CPU, for OpenVMS I64 systems. OpenVMS Alpha
Version 8.2 continues to provide capacity and per-disk licenses.
For OpenVMS I64 computers, a volume shadowing license is available
either as a separate product or as part of the collection of OpenVMS
products known as the Enterprise Operating Environment. For more
information about the Enterprise Operating Environment, refer to the
HP Operating Environments for OpenVMS Industry Standard 64 Version
8.2 for Integrity Servers Software Product Description (SPD
For more information about volume shadowing licensing,
refer to the HP Volume Shadowing for OpenVMS Software Product
Description (SPD 27.29.xx). For more information about the License
Management Facility, refer to the OpenVMS operating system SPD or the
HP OpenVMS License Management Utility Manual.
For more information about the HBMM new feature, refer to Chapter 6.
This chapter describes new features relating to application and system
programming in this version of the HP OpenVMS operating system.
4.1 Analyze Utility Enhancements (I64 Only)
The Analyze utility has been enhanced on OpenVMS I64 systems to analyze
Executable and Linkable Format (ELF) object and image files. Additional
information describing these enhancements has been added to the
HP OpenVMS DCL Dictionary for the ANALYZE/IMAGE and ANALYZE/OBJECT commands.
Using record formats and landmark values in the file, the ANALYZE
utility on OpenVMS I64 can determine the architecture type and whether
the file is an object or image file. Although ANALYZE accepts the
/OBJECT and /IMAGE qualifiers, these qualifiers do not restrict the
analysis to the specified file type.
For more information about the enhancements to the Analyze utility, see
the HP OpenVMS DCL Dictionary.
4.2 OpenVMS Calling Standard Changes for OpenVMS I64
The OpenVMS Calling Standard has been adapted for use on systems using
Intel Itanium processors and running OpenVMS I64.
The OpenVMS Calling Standard on the Intel Itanium processor family
is designed to follow the Itanium software conventions as much as
possible while avoiding user-visible differences from the OpenVMS VAX
and Alpha conventions. Changes to the Itanium convention were made
only where necessary to maintain compatibility with the historical
OpenVMS design. The goal was to minimize the cost and difficulty of
porting applications and OpenVMS itself to the Itanium
Refer to the HP OpenVMS Calling Standard for more information.
4.3 Checksum Utility
The Checksum utility calculates file, image, or object checksums for an
OpenVMS file. The CHECKSUM command invokes the utility, which can now
be run on I64, Alpha, or VAX platforms. The result, or checksum, is
available in the DCL symbol CHECKSUM$CHECKSUM.
For information about this utility, see the CHECKSUM command in the
HP OpenVMS DCL Dictionary.
4.3.1 CHECKSUM/OBJECT Enhanced for I64 Objects
CHECKSUM/OBJECT for I64 objects (ELF objects) provides new information
into the calculation for a checksum:
- EIDC (entity identification consistency check) information
- FPMODE (whole programming floating-point mode) information
If the EIDC or FPMODE fields are present in an object file, then the
I64 object checksum calculated by an earlier version of the Checksum
utility is different than the one calculated by Checksum on OpenVMS
The difference can be seen in the checksum of the ".note"
The previous version did not calculate any checksum for this section
while the new version calculates the checksum if the information is
present. Use the /SHOW=SECTIONS qualifier to view this behavior.
CHECKSUM/IMAGE is not affected; an image does not have EIDC or FPMODE
information in the ".note" section.
The file checksums (CHECKSUM without /IMAGE or /OBJECT) are also not
affected. File checksums use the whole file or the record structured
data to calculate the checksum.
4.4 C Run-Time Library Enhancements
The following sections describe the C Run-Time Library (RTL)
enhancements included in OpenVMS Version 8.2. These enhancements
provide improved UNIX portability, standards compliance, and the
flexibility of additional user-controlled feature selections. New C RTL
functions are also included. For more details, refer to the HP C
Run-Time Library Reference Manual for OpenVMS Systems.
4.4.1 File-Locking Functions
The following X/Open file-locking functions have been added. They allow
a user to lock and unlock files and provide access synchronization
across threaded programs:
4.4.2 Standard-Compliant stat Structure
An X/Open standard-compliant definition of the
structure and associated definitions are added. To use these new
definitions, applications must compile with a new feature macro defined:
The following new macros have also been added to support the
4.4.3 File-System Statistics Support
The following X/Open functions that return file system information have
been added in support of UNIX portability:
4.4.4 fcntl File Status Flags
The F_SETFL and F_GETFL command options are added to the
function to set and get file status flags.
4.4.5 UNIX Style Pipe Support
For added UNIX portability, the C RTL pipe implementation now uses
stream I/O as well as record I/O, under control of a new feature
The legacy behavior, record I/O, is the default behavior.
To enable stream I/O pipe support, define the DECC$STREAM_PIPE feature
logical name to ENABLE:
$ DEFINE DECC$STREAM_PIPE ENABLE
A pipe opened with the
function has its record attributes set to CR/LF carriage control
(fab$b_rat |= FAB$M_CR). UNIX systems do not insert CR/LF for pipes.
A new feature logical has been added to provide UNIX compatible
behavior. To override the default legacy behavior and thereby prevent
CR/LF carriage control from being added to the pipe records, define
DECC$POPEN_NO_CRLF_REC_ATTR to ENABLE:
$ DEFINE DECC$POPEN_NO_CRLF_REC_ATTR ENABLE
Be aware that enabling this feature might result in undesired behavior
from other functions, such as
, that rely on the carriage-return character.
4.4.7 glob and globfree 64-Bit Support
64-bit support is added for the
functions. As a result, the following additional function entry points
are now available for use with 32-bit and 64-bit pointer sizes,
TCP/IP socket routine has been added.
This routine is used for creating a pair of connected sockets and
requires the underlying TCP/IP product to have the TCPIP$SOCKETPAIR
4.5 DCE RPC Now Supports IEEE Floating-Point Type
DCE RPC for OpenVMS now supports both G_FLOAT and IEEE floating-point
types on OpenVMS Alpha and OpenVMS I64 platforms. The default
floating-point type on the Alpha platform remains G_FLOAT. The default
floating-point type on I64 platform is IEEE_FLOAT.
DCE RPC Application developers need to use the rpc_set_local_float_drep
call in their applications when using the non-default floating-point
DCE RPC on OpenVMS VAX platforms supports only G_FLOAT type.
4.6 OpenVMS Debugger
The following sections describe new features of the OpenVMS Debugger on
OpenVMS I64 systems.
4.6.1 Intel® Itanium® Hardware Support
OpenVMS I64 Debugger supports the following hardware registers:
- General registers R0 through R127
- Floating registers F0 through F127
- Branch registers B0 through B7
- Predicate registers P0 through P63. You can examine all predicate
register values using the symbol named PR.
- Application registers: AR16 (RSC), AR17 (BSP), AR18 (BSPSTORE),
AR19 (RNAT), AR25 (CSD), AR26 (SSD), AR32 (CCV), AR36 (UNAT), AR64
(PFS), AR65 (LC), AR66 (EC)
- A program counter named PC, synthesized from the hardware IP
register and the ri field of the PSR register
- Miscellaneous registers: CFM (current frame marker), UM (user
mask), PSP (previous stack pointer), and IIPA (previously executed
- Output register names OUT0 through OUT7. These names are provided
for convenience to make it easy to identify the registers that are used
to pass integer arguments from the current routine to a called routine.
For more information, refer to the HP OpenVMS Calling Standard.