HP OpenVMS Systems Documentation

HP OpenVMS System Management Utilities Reference Manual

This appendix describes OpenVMS system parameters.

System parameters can be grouped into categories, as shown in Section J.1.1. Also, each parameter can have one or more attributes, listed in Section J.1.1. Each parameter also has a value.

The parameters in this appendix are listed alphabetically along with their attributes.

J.1.1 Parameter Categories and Attributes

The system parameters can be divided into the following categories:

The user can also define four parameters: USERD1, USERD2, USER3, and USER4. USERD1 and USERD2 are dynamic.

Attributes for Parameters

Parameters can have one or more of the following attributes:

J.1.2 Values for Parameters

Default values for system parameters allow booting on any supported OpenVMS configuration. SYSGEN displays default values under the heading default when you enter the SYSGEN command SHOW [parameter-name] for one of the parameter categories or attributes. Reset the default parameter values with the USE DEFAULT command.

However, to avoid starting all layered products on a system that is not tuned for them, possibly causing the system to become nonoperational, set the STARTUP_P1 system parameter to "MIN."

The computed, installed value referred to in this section is the value derived by the AUTOGEN command procedure. (See the HP OpenVMS System Manager's Manual.)

J.2 Parameter Descriptions

This section describes system parameters and provides guidelines to help you decide whether you should consider modifying the parameters. The following attributes are indicated for the parameters:

AUTOGEN---A
DYNAMIC---D
FEEDBACK---F
GEN---G
MAJOR---M

This section alphabetically lists and describes the system parameters in all categories.

Parameters

ACP_BASEPRIO (D)

ACP_BASEPRIO sets the base priority for all ACPs. The DCL command SET PROCESS/PRIORITY can be used to reset the base priorities of individual ACPs. ACP_BASEPRIO is not applicable for XQPs.

ACP_DATACHECK (D)

ACP_DATACHECK controls the consistency checks that are performed on internal file system metadata such as file headers.

ACP_DATACHECK is a bit mask. The following table shows the bits that are defined currently:

Bit	Description
0	Set this bit to perform consistency checks on read operations. When this bit is set, the IO$M_DATACHECK function modifier is automatically set on all subsequent IO$_READLBLK operations that read file system metadata (see the HP OpenVMS I/O User's Reference Manual).
1	Set this bit to perform consistency checks on write operations. When this bit is set, the IO$M_DATACHECK function modifier is automatically set on all subsequent IO$_WRITELBLK operations that read file system metadata (see the HP OpenVMS I/O User's Reference Manual).
2	Set this bit to perform read-after-write consistency checks. This is similar to setting bit 1, except that in this case the file system does the checks, not the lower level device or disk driver. Note that read-after-write consistency checks are not allowed on deferred writes. Deferred writes are turned off if this bit is set.
3	Reserved for HP use only; must be zero.
4	Reserved for HP use only; must be zero.
5 and 6	These two bits control the checks that are performed on reads and writes of directory blocks. You can select one of four different levels: To Check That... Select This Level... By Setting Bit 6 to... And Bit 5 to... The block is a valid directory block (reads only) 0 0 0 The block is a valid directory block (reads and writes) 1 0 1 The block is a valid directory block and contains valid entries (reads and writes) 2 1 0 The block is a valid directory block and contains valid entries in correct alphanumeric order (reads and writes) 3 1 1 When you set the SYSTEM_CHECK system parameter to 1, you enable level 3 checking of directory blocks. Write errors result in BUGCHECK and crash your system; read errors exit with error status SS$_BADDIRECTORY.
7	Reserved for HP use only; must be zero.

ACP_DINDXCACHE (A,D,F)

ACP_DINDXCACHE controls the size of the directory index cache and the number of buffers used on a cachewide basis. Also, ACP_DINDXCACHE builds a temporary index into the directory file, thereby reducing search time and directory header lookup operations.

ACP_DIRCACHE (A,D,F)

ACP_DIRCACHE sets the number of pages for caching directory blocks. Too small a value causes excessive XQP I/O operations, while too large a value causes excessive physical memory to be consumed by the directory data block cache.

ACP_EXTCACHE (D,F)

ACP_EXTCACHE sets the number of entries in the extent cache. Each entry points to one contiguous area of free space on disk. A specification of 0 means no cache. Too small a value causes excessive XQP I/O operations, while too large a value causes excessive physical memory to be consumed by the extent cache.

ACP_EXTLIMIT (D)

ACP_EXTLIMIT specifies the maximum amount of free space to which the extent cache can point, expressed in thousandths of the currently available free blocks on the disk. For example, if available free space on the disk is 20,000 blocks, a specification of 10 limits the extent cache to 200 blocks.

The computed, installed value is usually adequate. Users with four or more OpenVMS Cluster node systems might want to adjust this parameter.

ACP_FIDCACHE (D,F)

ACP_FIDCACHE sets the number of file identification slots cached. A specification of 1 means no cache. Too small a value causes excessive XQP I/O operations, while too large a value causes excessive physical memory to be consumed by the FID caches.

ACP_HDRCACHE (A,D,F)

ACP_HDRCACHE sets the number of pages for caching file header blocks. Too small a value causes excessive XQP I/O operations, while too large a value causes excessive physical memory to be consumed by the file header caches.

ACP_MAPCACHE (A,D,F)

ACP_MAPCACHE sets the number of pages for caching index file bitmap blocks. Too small a value causes excessive XQP I/O operations, while too large a value causes excessive physical memory to be consumed by the bitmap cache.

ACP_MAXREAD (D)

ACP_MAXREAD sets the maximum number of directory blocks read in one I/O operation.

ACP_MULTIPLE (A,D)

ACP_MULTIPLE enables (1) or disables (0) the default creation of a separate disk XQP cache for each volume mounted on a different device type. Prior to Version 4.0, a separate ACP process was created for each device type if this parameter was enabled. Because ACP operations are now handled by the per process XQP, such separate processes are no longer created. In general, having multiple caches is unnecessary. One large cache is more efficient than several small ones. ACP_MULTIPLE can be overridden on an individual-volume basis with the DCL command MOUNT.

ACP_QUOCACHE (A,D,F)

ACP_QUOCACHE sets the number of quota file entries cached. A specification of 0 means no cache. Too small a value causes excessive XQP I/O operations, while too large a value causes excessive physical memory to be consumed by the quota caches.

ACP_REBLDSYSD

ACP_REBLDSYSD specifies whether the system disk should be rebuilt if it was improperly dismounted with extent caching, file number caching, or disk quota caching enabled. The ACP_REBLDSYSD default value (1) ensures that the system disk is rebuilt. Setting the value to 0 means the disk is not rebuilt.

Depending on the amount of caching enabled on the volume before it was dismounted, the rebuild operation may consume a considerable amount of time. Setting the value of ACP_REBLDSYSD to 0 specifies that the disk should be returned to active service immediately. If you set ACP_REBLDSYSD to 0, you can enter the DCL command SET VOLUME/REBUILD at any time to rebuild the disk.

ACP_SHARE (D)

ACP_SHARE enables (0) or disables (1) the creation of a global section for the first ACP used, enabling succeeding ACPs to share its code. This parameter should be set to 0 when ACP_MULTIPLE is on.

ACP_SWAPFLGS (A,D)

ACP_SWAPFLGS enables or disables swap through the value of a 4-bit number for the following four classes of ACPs:

Bit	Class of ACP
0	Disks mounted by MOUNT/SYSTEM
1	Disks mounted by MOUNT/GROUP
2	Private disks
3	Magnetic tape ACP

If the value of the bit is 1, the corresponding class of ACPs can be swapped. The value of decimal 15 (hexadecimal F---all bits on) enables swap for all classes of ACP. A value of decimal 14 disables swap for ACPs for volumes mounted with the /SYSTEM qualifier but leaves swap enabled for all other ACPs. Note that one has only disk ACPs present if they are specifically requested at mount time or if a Files-11 On-Disk Structure Level 1 disk is mounted. In general, only bit 3 is significant because usually no file ACPs exist.

ACP_SYSACC (A,D)

ACP_SYSACC sets the number of directory file control blocks (FCBs) that are cached for disks mounted with the /SYSTEM qualifier. Each directory FCB contains a 16-byte array containing the first letter of the last entry in each block of the directory (or group of blocks if the directory exceeds 16 blocks). Since entries in a directory are alphabetical, the cached FCB provides quick access to a required directory block. This parameter value should be roughly equivalent to the number of directories that are in use concurrently on each system volume. It might be overridden on a per-volume basis with the /ACCESSED qualifier to the DCL command MOUNT. The value should be kept low in systems with small physical memory and little file activity, because the FCBs require a significant amount of space in the nonpaged dynamic pool.

Too small a value causes excessive XQP I/O operations, while too large a value causes excessive physical memory to be consumed by the FCB caches.

ACP_WINDOW (D)

ACP_WINDOW sets the default number of window pointers to be allocated in a window for a default file access, for disks mounted with the /SYSTEM qualifier.

ACP_WORKSET (D)

ACP_WORKSET sets the default size of a working set for an ACP. A specification of 0 permits the ACP to calculate the size. This value should be nonzero only on small systems where memory is tight. Too small a value causes excessive ACP page, while too large a value causes excessive physical memory to be consumed by the ACP. Note that this parameter has no effect on the per-process XQP.

ACP_WRITEBACK (D)

ACP_WRITEBACK is a dynamic system parameter that controls whether deferred writes to file headers are enabled. The default value is 1, which enables deferred writes to file headers. To disable the feature, set ACP_WRITEBACK to 0.

This system parameter affects only applications like PATHWORKS that can request deferred writes to file headers. Note that the deferred write feature is not available on Files-11 ODS--1 volumes.

ACP_XQP_RES

ACP_XQP_RES controls whether the XQP is currently in memory. The default value (1) specifies that the XQP is permanently in memory. Change the default only on restricted memory systems with a small number of users and little or no file activity that requires XQP intervention. Such activity includes file opens, closes, directory lookups, and window turns.

AFFINITY_SKIP

AFFINITY_SKIP controls the breaking of implicit affinity. The value indicates the number of times a process is skipped before being moved.

This special parameter is used by HP and is subject to change. Do not change this parameter unless HP recommends that you do so.

AFFINITY_TIME

AFFINITY_TIME controls the breaking of implicit affinity. The value indicates how long a process remains on the compute queue.

This special parameter is used by HP and is subject to change. Do not change this parameter unless HP recommends that you do so.

ALLOCLASS

ALLOCLASS determines the device allocation class for the system. The device allocation class is used to derive a common lock resource name for multiple access paths to the same device.

ARB_SUPPORT (D)

(Alpha and I64) The Access Rights Block (ARB) compatibility option, the ARB_SUPPORT system parameter, is provided specifically to support products that have not yet been updated to use the new per-thread security Persona Security Block (PSB) data structure instead of the ARB. Changing the value of ARB_SUPPORT from 2 or 3 (the default) to any other value can affect the operation of these products.

Note HP recommends that all Version 7.3-1 systems have the ARB_SUPPORT parameter set to 3 (the default). Do not change the ARB_SUPPORT parameter to any other value until all products dependent on the ARB and associated structures have been modified for the new environment.

The following table describes ARB_SUPPORT parameters:

ARB_SUPPORT Parameter	Value	Behavior
ISS$C_ARB_NONE	0	The obsolete kernel data cells are not maintained by the system. Fields are initialized to zero (or set to invalid pointers) at process creation.
ISS$C_ARB_CLEAR	1	The obsolete kernel data cells are cleared (or set to invalid pointers) when the code would have set up values for backward compatibility.
ISS$C_ARB_READ_ONLY	2	The obsolete cells are updated with corresponding security information stored in the current PSB when a $PERSONA_ASSUME is issued.
ISS$C_ARB_FULL	3 (default)	Data is moved from the obsolete cells to the currently active PSB on any security-based operation.

AUTO_DLIGHT_SAV

AUTO_DLIGHT_SAV is set to either 1 or 0. The default is 0.

If AUTO_DLIGHT_SAV is set to 1, OpenVMS automatically makes the change to and from daylight saving time.

AWSMIN (D)

On VAX systems, AWSMIN establishes the lowest number of pages to which a working set limit can be decreased by automatic working set adjustment.

On Alpha and I64 systems, AWSMIN establishes the lowest number of pagelets to which a working set limit can be decreased by automatic adjustment of the working set.

AWSTIME (D)

AWSTIME specifies the minimum amount of processor time that must elapse for the system to collect a significant sample of a working set's page fault rate. The time is expressed in units of 10 milliseconds. The default value of 20, for example, is 200 milliseconds.

Some application configurations that have a large number of memory-intensive processes may benefit if the value is reduced. The value can be as low as 4.

AWSTIME expiration is checked only at quantum end. Reducing its value and not reducing QUANTUM effectively sets the value of AWSTIME equal to the value of QUANTUM.

BALSETCNT (A,G,D,M)

BALSETCNT sets the number of balance set slots in the system page table. Each memory-resident working set requires one balance set slot.

You can monitor the active system with the DCL command SHOW MEMORY or the MONITOR PROCESSES command of the Monitor utility to determine the actual maximum number of working sets in memory. If this number is significantly lower than the value of BALSETCNT, this parameter value could be lowered. If all balance set slots are being used, raise the value of BALSETCNT.

Never set BALSETCNT to a value higher than 2 less than MAXPROCESSCNT. If physical memory is a significant system constraint, consider lowering this value even further. However, if your system runs with a number of processes nearly equal to MAXPROCESSCNT, lowering BALSETCNT forces swapping to occur, which can affect system performance.

BALSETCNT is no longer a strict setting of the number of processes that might be resident in memory. The swapper tries to reduce the number of resident processes down to BALSETCNT. However, if the total number of active processes and processes that have disabled swapping exceeds BALSETCNT, the swapper does not force processes out of memory just to meet the BALSETCNT setting.

Note that on VAX systems, virtual balance slots (VBS) can affect the values of BALSETCNT and MAXPROCESSCNT.

BORROWLIM (A,D,M)

BORROWLIM defines the minimum number of pages required on the free-page list before the system permits process growth beyond the working set quota (WSQUOTA) for the process. This parameter should always be greater than FREELIM.

This parameter allows a process to grow beyond the value set by the working set quota (WSQUOTA) to the working set quota extent (WSEXTENT) on a system that has a substantial memory on the free-page list. This automatic working set adjustment also depends upon the values of parameters WSINC, PFRATH, and AWSTIME.

Working set growth attempts to alleviate heavy page faulting. To make use of this growth, you must also set the user's WSEXTENT authorization quota to a larger number than the WSQUOTA value.

BREAKPOINTS

If XDELTA is loaded, BREAKPOINTS enables additional built-in calls for XDELTA during the boot sequence. The breakpoints that are enabled may change from release to release of OpenVMS.

This special parameter is used by HP and is subject to change. Do not change this parameter unless HP recommends that you do so.

BUGCHECKFATAL (D)

BUGCHECKFATAL enables or disables the conversion of nonfatal bugchecks into fatal bugchecks. The system must be rebooted on a fatal bugcheck. A nonfatal bugcheck places an entry only in the error log and deletes the corresponding process.

This parameter should normally be OFF (0); you should set it ON (1) only when the executive is being debugged.

Setting the SYSTEM_CHECK parameter to 1 has the effect of setting BUGCHECKFATAL to ON (1).

BUGREBOOT (D)

BUGREBOOT enables or disables automatic rebooting of the system if a fatal bugcheck occurs. This parameter should normally be on (1); set it off (0) only when the executive is being debugged.

CHANNELCNT

CHANNELCNT specifies the maximum number of I/O channels available to processes and to the system. The FILLM quota can be used to reduce the maximum number of I/O channels for a process. A process with a FILLM quota larger than CHANNELCNT is nevertheless limited to the maximum number of I/O channels specified by CHANNELCNT.

CHECK_CLUSTER

(VAX only) CHECK_CLUSTER is the VAXCLUSTER parameter sanity check. When CHECK_CLUSTER is set to 1, SYSBOOT outputs a warning message and forces a conversational boot if it detects that the VAXCLUSTER parameter is set to 0.

CLASS_PROT (D)

CLASS_PROT performs the nondiscretionary classification checks. CLASS_PROT is also checked by XQP to determine if a classification block should be added to the header of any created files.

CLISYMTBL (D)

CLISYMTBL sets the size of the command interpreter symbol table, which controls the number of DCL symbols that can be created.

CLOCK_INTERVAL

(VAX only) CLOCK_INTERVAL sets the number of microseconds between the hardware interval timer clock interrupts. It has no effect on processors that have implemented only the subset interval clock registers.

This special parameter is used by HP and is subject to change. Do not change this parameter unless HP recommends that you do so.

CLUSTER_CREDITS

CLUSTER_CREDITS specifies the number of per-connection buffers a node allocates to receiving VMS$VAXcluster communications.

If the SHOW CLUSTER command displays a high number of credit waits for the VMS$VAXcluster connection, you might consider increasing the value of CLUSTER_CREDITS on the other node. However, in large cluster configurations, setting this value unnecessarily high consumes a large quantity of nonpaged pool. Each receive buffer is at least SCSMAXMSG bytes in size but might be substantially larger depending on the underlying transport.

It is not required for all nodes in the cluster to have the same value for CLUSTER_CREDITS.

The default value is currently 32. Unless a system has very constrained memory available, HP recommends that these values not be increased.

CONCEAL_DEVICES

CONCEAL_DEVICES enables or disables the use of concealed devices. By default, this parameter is set to enable concealed devices (1).

This special parameter is used by HP and is subject to change. Do not change this parameter unless HP recommends that you do so.

CRD_CONTROL

This special parameter is used by HP and is subject to change. Do not change this parameter unless HP recommends that you do so.

---

Previous

Next

Contents

Index