Previous Topic: Parameter SummaryNext Topic: Using the CommandFacility as a Subprogram


Parameter Descriptions

ABEND_ON_DEADLOCK

Forces the abnormal termination of a task that encounters a database resource deadlock. In normal CV operations, a database resource deadlock results in control being returned to the application program with an indication that a deadlock occurred. This parameter causes the task to be abended instead.

Note: It is meaningful only in the SYSIDMS file associated with a central version.

ABEND_SVC_DUMP=ON|OFF

Causes an SVC dump to be requested when CV abends prematurely (ON) or not at all (OFF).

Default: ON

Note: Specifying OFF is not recommended. If the CV address space abends, a dump will be required to determine the cause of the abend.

ABENDTRACE=ON|OFF

Activates the tracing of various pieces of CA IDMS data when using CA Optimizer/II or CA SymDump Batch. This parameter is meaningful only in the SYSIDMS file associated with a batch job.

Note: You must have CA Optimizer/II or CA SymDump Batch installed to use this parameter. However, if CA Optimizer/II or CA SymDump Batch is installed and you do not specify ABENDTRACE=ON, you can still get an abbreviated trace of CA IDMS activity by setting optional bit 265 in RHDCOPTF.

ABENDTRACE_ENTRIES=nnn

Overrides the default number of entries being traced by ABENDTRACE. This parameter is meaningful only in the SYSIDMS file associated with a batch job.

Note: You must have CA Optimizer/II or CA SymDump Batch installed to use this parameter.

ABENDTRACE_SUBSCHEMA_DISPLAY=ON

Activates the display of information from the subschema in use at the time of abend when using ABENDTRACE. This parameter is meaningful only in the SYSIDMS file associated with a batch job.

Note: You must have CA Optimizer/II or CA SymDump Batch installed to use this parameter.

ABENDTRACE_VIBSNAP=ON

Causes the dump of the VIB at the time of abend when using ABENDTRACE. This parameter is meaningful only in the SYSIDMS file associated with a batch job.

Note: You must have CA Optimizer/II or CA SymDump Batch installed to use this parameter.

ADJUNCT_TRACE_TABLE=

table-size

KB|MB

Specifies the size of the adjunct trace table in kilobytes (KB) or megabytes (MB).

Limit: 0–9999

Default: 0

Note: Any non-zero adjunct table size established by a SYSIDMS parameter overrides the adjunct trace table size specified in the system definition.

ARCHIVE_JOURNAL_WARNING_PERCENT=percent-number

Specifies the threshold percent that ARCHIVE JOURNAL uses to issue a warning message that a journal file is nearly full. If the amount of available space after condensing a journal is less than percent-number, ARCHIVE JOURNAL issues a warning message indicating that the journal is nearly full of condensed segments.

Default: 10 percent

AREA_VALIDATION_MSGS=ON|OFF

ON causes the informational messages DB347042 and DB347043 to be displayed on the JES log during startup and shutdown for each area being shared in a SYSPLEX data sharing environment. If you are sharing many areas this can cause the JES log to be congested.

Default: OFF

Note: This parameter is only applicable in a SYSPLEX data sharing environment.

BCF_INPUT_80=ON|OFF

IDMSBCF input is done over 80 characters by a #LINEIN INAREA=CARD,MAXIN=80,... statement. However, afterwards all information starting from column 73 until column 80 is cleared and overwritten by spaces. That is, the actual maximum input size is only 72 characters.

Executing any SQL PUNCH command may produce a SYSPCH output exceeding column 72. When using this output as input for an IDMSBCF job afterwards, may result in possible syntax errors because the input is truncated after column 72.

By specifying SYSIDMS parameter BCF_INPUT_80=ON, information between columns 73 and 80 is not cleared (that is, input for BCF will be over 80 characters). In all other cases (that is, no BCF_INPUT_80 parameter or not specifying BCF_INPUT_80=OFF), it defaults to 72 characters.

Default: OFF

BUFFER_PURGE

Causes updated pages to be written to the database whenever the number of buffers containing such pages exceeds 1/4 of the number of pages in the buffer pool. This parameter may improve the performance of local mode update jobs that do not issue frequent COMMITs, since it will make buffers available for the use of pre-fetch. It has meaning only for local mode batch jobs.

BUFFERSTAT

Produces a report containing buffer pool I/O statistics that can be used for tuning purposes. The report will be written to SYSLST at the end of the job. This parameter has meaning only for local mode batch jobs.

Note: For a description of the fields in the report, see the BUFFERSTAT Report Field Descriptions.

CICS_NAME=cics-name

Specifies a 1 to 4 character value that identifies the CICS system name being started.

CVMACH=cms-machine-name

(z/VM only) Specifies the virtual machine in which the DC/UCF system is executing.

CVNUM=nnn

(z/VM only) Specifies the number of the central version that is accessible by CMS and is used to route database requests through the IDMSVMCF facility.

Limits: 0-255

CVRETRY=ON|OFF

ON indicates that the following message is displayed on the operator console when the CA IDMS central version is not active:

CV nnn NOT ACTIVE. REPLY RETRY OR CANCEL.

Default: ON

CVRETRY_MSG_CODES=

descriptor-route-codes

Specifies the descriptor and route codes to be used for batch message DC208002 (CV cv-number NOT ACTIVE. REPLY RETRY OR CANCEL). descriptor-route-codes must be an eight-digit hexadecimal value.

The first four digits of descriptor-route-codes represent the descriptor codes and the last four digits represent the route codes. Each bit within the descriptor or route codes represents a code value. The first bit (x’8000’) represents code value 1 and the last bit (x’0001’) represents code value 16. Multiple bits can be on in each set of codes so that x’8101’ represents code values 1, 8 and 16.

Default: 00004000 (representing descriptor code zero (0) and route code two)

Note: descriptor-route-codes value is automatically considered to be in the hexadecimal format.

CV_STARTUP_XA_REGION_MB=nnn

Specifies the size of the initial XA storage pool acquired during early CV startup.

nnn

Specifies the size in MB (megabytes) of the initial XA storage pool.

Default: 32 MB

DB_DEADLOCK_DUMP

Specifies that a dump will be produced for a task that is abended due to a database resource deadlock. This parameter is used in conjunction with the ABEND_ON_DEADLOCK parameter. If not specified, no dump will be produced when a task is abended due to a database deadlock.

DBIO_HICCUP=nn

By default, the CV will generate a hiccup wait if a task accesses 100 pages without generating a wait. This parameter allows 1-99 pages to be accessed before a wait is generated. Specifying a value greater than 100 results in a page count of 100.

Note: It is recommended that you do not use very low values.

DBAN_SORT_PASSES=sort-passes

Specifies the number of sorts that IDMSDBAN uses when auditing index orphan chains.

Default: 1

The higher the number of sorts, the greater the number of errors that can be detected, if any exist. However, increasing the number of sorts also increases the overhead for auditing orphan chains, which can be significant when processing large indexes. Because most errors are detected with a single sort, the default of one is sufficient in most cases. You should increase the number of sorts beyond 1 only when you need to identify every error in an index.

DBNAME=database-name

For non-SQL applications, specifies the name of the database to access at runtime. database-name is either a segment name or a DBNAME defined in a database name table. For SQL applications, it has no impact.

DC_DEADLOCK_NODUMP

Specifies that a dump is not produced for a task that is abended due to a DC resource deadlock. This parameter overrides the DUMP/NODUMP sysgen parameter.

DC_DEADLOCK_0029

Specifies that tasks that encounter a DC resource deadlock be abended with a code of 0029 rather than a code of DEAD.

DC_SCRATCH=ON|OFF

ON allows local jobs to use the central version's scratch area (DDLDCSCR) when a local scratch area (DDLOCSCR) is not defined in the DMCL.

Default: ON

DCNAME=member-name

(z/OS only) Specifies the member name of the system within a data sharing group. This name also becomes the system (node) name, overriding the value specified in the system definition. member-name must consist of characters A-Z, 0-9, $, #, or @.

Limits: 1-8 characters

Note: This parameter is only applicable in a data sharing environment.

DEADLOCK_ABEND_ERUS

Specifies that ERUS tasks that encounter a database resource deadlock be abnormally terminated. This parameter is meaningful only if the ABEND_ON_DEADLOCK parameter is also specified.

DEADLOCK_ABEND_0029

Specifies that tasks that are abended due to database resource deadlocks use a code of 0029 rather xx29, where "xx" represents the major code of the DML request that was being issued at the time of the abend. This parameter has meaning only if the ABEND_ON_DEADLOCK parameter is also specified.

DEADLOCK_DETAILS=ON|OFF

ON specifies that more detail be provided in a deadlock situation.

Default: OFF

DICTNAME=dictionary-name

Specifies a dictionary to use when loading a subschema from a load area. For dictionary-related tools like CA IDMS compilers and pre-compilers, IDMSBCF, and so on, dictionary-name specifies the dictionary to access at run time. For SQL applications, dictionary-name specifies the name of the dictionary to connect to at run time.

DICTNODE=dictionary-node-name

For SQL applications and dictionary-related tools under the central version, specifies the name of the DC/UCF system that controls the dictionary to access at run time. For applications running in local mode, this parameter is not applicable.

DISABLE_SVC_SCREEN=SVC-number

Specifies the number of an SVC for which screening is to be disabled.

Limits: 0–255

Disabling screening of an SVC allows it to be issued by a program executing within the DC/UCF address space.

The DISABLE_SVC_SCREEN parameter can be specified multiple times.

Note: Use this parameter with caution, since issuing unscreened SVCs within the DC/UCF system can degrade performance and result in abends. For more information about SVC screening, see DML Reference Guide for Assembler.

DLBLMOD=ON|OFF

(z/VSE only) ON specifies that the DLBL type in the disk label will be changed from 'DA' to 'SD' when sequential processing (IDMSQSAM) is activated. After the disk labels are processed as 'SD' during the QSAM file OPEN, the DLBLs are changed back to 'DA' to allow random BDAM processing.

Default: OFF

DMCL=dmcl-name

Specifies the name of the DMCL load module to use in local mode.

Default: IDMSDMCL

DMLTRACE=ON|OFF

ON activates a trace facility that traces all navigational DML requests made by an application.

Default: OFF

DSGROUP=data-sharing-group-name

(z/OS only) Specifies the name of the data sharing group of which this system is a member. All CA IDMS systems that are members of the same group must specify the same group name. The data-sharing-group-name must consist of characters A-Z, 0-9, $, #, or @. Names that begin with SYS or UNDESIG are reserved and cannot be used. Names that begin with A-I may be in use by the operating system and should be avoided.

Limits: 1-8 characters

Note: This parameter is only applicable in a data sharing environment.

DYNALLOC_WAIT=ON|OFF

ON specifies that dynamic allocation will do an ENQ wait for the DSN until it becomes available, when a data set is in use.

OFF specifies that when DYNALLOC_WAIT_SECONDS option is zero or not specified, the dynamic allocation request will fail.

This option applies to local mode jobs and to CV startup.

Default: OFF

DYNALLOC_WAIT_SECONDS=nnn

Specifies that when a data set is in use, dynamic allocation will wait for a specified interval and retry the allocation. If the DSN is still unavailable it will repeat the process until the data set is available. If this option is specified with a non-zero value, it will override the DYNALLOC_WAIT option. If zero is specified and the DYNALLOC_WAIT option is OFF or not specified; the dynamic allocation request will fail.

This option applies to local mode jobs and to CV startup.

Limits: 255

ECHO=ON|OFF

Indicates whether SYSIDMS parameters are displayed on the JES log.

Default: OFF

ENABLE_RRS=ON|OFF

(z/OS only) ON activates RRS support for batch applications.

Default: OFF

EVAL_BASE_YEAR=base-year

Specifies the base year to be assumed by the DATEDIFF and DATEOFF built-in functions. The base year is used to determine whether a two-digit year is considered to be in the twentieth or twenty-first centuries. A year whose value is greater than the base year is considered to be in the twentieth century; values less than or equal to the base year are considered to be in the twenty-first century.

Limits: 1 - 99

Default: 68

EVAL_CENTURY_VALIDATION=ON|OFF

Specifies whether century values are to be validated by built-in functions that accept four-digit years such as GOODDATEX.

ON

Validate century values.

OFF

Does not validate century values.

Default: OFF

EVAL_LOW_CENTURY=low-century

Specifies the lowest century value to be considered valid. The value specified is used during century validation in built-in functions that accept four-digit years such as GOODDATEX. Centuries are validated only if the EVAL CENTURY VALIDATION is ON.

Limits: 1 - 99

Default: 19

EVAL_HIGH_CENTURY=high-century

Specifies the highest century value to be considered valid. The value specified is used during century validation in built-in functions that accept 4-digit years such as GOODDATEX. Centuries are validated only if the EVAL CENTURY VALIDATION is ON.

Limits: 1 - 99

Default: 20

EXCLUDE_TCP/IP_STACK=stack=name

(z/OS and z/VM only) Creates an EXCLUDE list of up to eight TCP/IP stack names to override the list of stacks supplied by the operating system (z/OS) or through SYSGEN (and z/VM).

All the stack-names that are specified on these SYSIDMS parameters are excluded from the runtime list of stacks. INCLUDE_TCP/IP_STACK and EXCLUDE_TCP/IP_STACK are mutually exclusive parameters and support wildcards.

FILE_BUF=ddname=nnnnn

Allows users to increase the number of pages in a buffer used by a specific file for a local mode job without having to change the DMCL. In CV, a DCMT command can be used to alter the number of pages in a buffer. The FILE_BUF parameter provides a similar capability for local mode jobs. If specified, the number of pages in the buffer pool associated with the specified file is increased by nnnnn.

This parameter can be used to tune PREFETCH processing by allowing the local mode user to increase the number of pages in a specific buffer for a job and thereby maximize the benefit of prefetch processing.

IDMSQSAM=ON|OFF

ON activates the IDMSQSAM facility (sequential access for look-ahead database reads).

Default: OFF

IGNORE_SYSTRK_DMCL=ON|OFF

Specifies whether to disable building the runtime DMCL from the SYSTRK file and instead force the use of the DMCL load module.

ON

Specifies to disable use of SYSTRK for building the runtime DMCL.

OFF

Specifies to build the runtime DMCL from SYSTRK if the CV was not previously shutdown normally.

Default: OFF

INCLUDE_TCP/IP_STACK=stack-name

(z/OS and z/VM only) Creates an INCLUDE list of up to eight TCP/IP stack names to override the list of stacks supplied by the operating system (z/OS) or through SYSGEN (z/OS and z/VM). For z/VM only, it can be used to include a new stack.

All the stack-names that are not specified on these SYSIDMS parameters are excluded from the runtime list of stacks. INCLUDE_TCP/IP STACK and EXCLUDE_TCP/IP_STACK are mutually exclusive parameters and support wildcards.

JOURNAL=ON|OFF

Specifies whether journaling will be performed in local mode. OFF specifies that local mode journaling will not be performed, even if there are tape journals defined in the DMCL.

Default: ON

JRNLDTS=yyyy-mm-dd-hh.mm.ss.nnnnnn

This parameter provides a way to bypass a date time stamp mismatch problem between the DMCL and the journal files. The yyy-mm-dd-hh.mm.ss.nnnnnn is the date time stamp on the journal file. This should only be used if you know that the reason for the mismatch will not cause a problem. Inappropriate use of this parameter may cause database corruption.

LANG=xxxxxxxxxxxxxxxxxxx

Sets an alternate environment for DBCS support. This parameter is useful for local mode batch jobs and is equivalent to issuing the DCUF SET LANG command for online users.

Limits: 19 characters

LENGTH_PAGE=nnn

Specifies the maximum number of lines to be printed on a page.

Limits: 10-32,767 lines

Default: 60

LIMIT_PREFETCH=nnn

When a batch job is performing an area sweep, prefetch takes 3/4 of the number of buffers and divides by the number of records per track to determine the maximum number of start I/Os it can issue. If the number is greater than 100, a value of 100 is used. At the start of a job this can swamp the system with I/O requests.

By coding LIMIT_PREFETCH=nnn prefetch will only issue nnn I/Os at any time. If this number is greater than the value calculated by CA IDMS, the lower value will be used.

LOADAREA=ON|OFF

Specifies whether the dictionary load (DDLDCLOD) area is to be accessed when loading a module. If OFF is specified, the dictionary load area will not be accessed. You should specify OFF only when all load modules are linked into an OPSYS load library.

Default: ON

LOCAL=ON|OFF

Specifies whether a batch job is to execute in local mode. If ON is specified, all requests are processed locally even if an IDMSOPTI is link-edited with the program, or a SYSCTL file is specified in the JCL.

Default: OFF

LOCAL_DYNAMIC_ALLOCATION= ON|OFF

OFF directs a local CA IDMS batch job to ignore any DSN information defined in the DMCL for database files, and requires that the DSN information be included in the JCL in order to access a database file.

Default: ON

LOCAL_NOJOURNAL_RETRIEVAL

Specifies that local batch jobs not journal RETRIEVAL ONLY transactions.

LOCALPUR=ON|OFF

ON forces the purging of the local mode buffer pool whenever a transaction terminates.

This parameter addresses a change in the way local mode buffers are handled (between 10.21 and later releases). In release 10.21 a local mode job that had multiple transactions (run units) would have separate buffer pools for each transaction (and each transaction would have no knowledge of the others). When a transaction terminated its buffer pool would be purged. Starting in release 12.0, a local job with multiple transactions will have just one buffer pool shared by all transactions. When a transaction terminates the buffer remains unchanged until the last transaction terminates at which time the shared buffer pool is purged. To make the system purge the common buffer pool when each transaction terminates (and therefore mimic what happened in release 10.21), use the parm LOCALPUR=ON.

Default: OFF

Note: This parameter should only be specified if a compatibility problem is encountered, since there can be performance implications in specifying LOCALPUR=ON.

MSGDICT=ON|OFF

Specifies whether the dictionary message (SYSMSG) area is to be accessed in order to retrieve the text of messages. If OFF is specified, the dictionary message area is not accessed. OFF should be specified only when using a DMCL that does not contain the SYSMSG segment, such as during installation.

Default: ON

MULTIDSN=ON|OFF

(z/VSE only) ON specifies that tape files may span multiple volumes. At the end of a tape reel, EOF (end of file) or EOV (end of volume) prompts the user to specify an END OF JOB or an END OF VOLUME condition. OFF specifies that END OF JOB is automatically the condition at the end of a tape reel.

Default: OFF

NODENAME=nodename

For non-SQL applications running under the central version, identifies the DC/UCF system to bind to at run time.

PARM='parameter-string'

Allows you to specify parameters typically specified in a JCL EXEC PARM statement. The format is the same as the IBM PARM parameter on the EXEC JCL statement. parameter-string can contain any 1 through 256 character parameter and can be specified on multiple lines.

PDAT_LOCATION=ANY|BELOW

Specifies whether the UCF line driver allocates PDAT storage above or below the 16 MB line. ANY, the default, allocates PDAT storage above the line. BELOW allocates PDAT storage below the line. BELOW must be specified if a back-end PDAT application runs with AMODE 24.

Default: ANY

DATE_SIMULATOR_SVC=nnn

Causes the CA IDMS CV or batch job to issue an SVC to call the DATE SIMULATOR and replaces the STCK (STORE CLOCK) commands as documented in the DATE SIMULATOR instructions provided by the Date Simulator vendor.

The purpose of this parameter is to allow clients to run their DATE SIMULATOR products without zapping the CA IDMS modules.

PREFETCH=ON|OFF

OFF overrides the default ON and prevents CA IDMS from prefetching database pages, the normal processing when an area or index sweep is detected. Specify OFF for a local batch job to prevent prefetching database pages for the job step. Specify OFF in the SYSIDMS file associated with a central version to prevent prefetching pages for all transactions running with the central version.

Default: ON

PREFETCH_BUF=nnnnn

Specifies the minimum number of pages in a buffer pool that must be present before CA IDMS will use prefetch processing for non-area sweep requests. This parameter applies to both local and central version environments.

PRMPT_MSG='prompt_message'

Customizes the WTOR (Write-To-Operator-With-Reply) message. Customizing this message lets you distinguish the different CVs that are running.

Enclose prompt_message in single (') or double (") quotes. You can write the message in free-form, but do not enter internal quotes.

Limits: The message can be 1 to 58 characters long (without quotes).

Default: If the parameter is not specified, the following message is displayed:

REPLY WITH REQUEST TO IDMS Vnnn 

nnn is the system number, which is present regardless if the default or a customized message is displayed.

PROCTRACE=ON|OFF

ON activates a trace of key user blocks that participate in an SQL PROCEDURE call.

Default: OFF

QSAMAREA=qsam-area-name

Specifies the physical area in the DMCL for which the IDMSQSAM facility will do look-ahead reads. If this parameter is omitted and the IDMSQSAM=ON parameter is specified, the look-ahead reads will be performed on the first area that is accessed by the transaction.

Note: This parameter may also be used against journal files when executing the ARCHIVE JOURNAL utility by specifying QSAMAREA=ARCHIVE.JOURNAL in the utility's SYSIDMS file.

QSAMBUF#=nnn

(z/OS only) Specifies the number of buffers to use when the IDMSQSAM facility is active. QSAMBUF# enables you to set the number of QSAM buffers to be used without having to code JCL for the file being processed by IDMSQSAM.

If QSAMBUF# is not specified, the number of buffers is determined by the DCB parameter BUFNO=nnn, or defaults to 5 buffers.

Limits: 1-255

Default: 5 buffers

QSAMTRACE=ON|OFF

ON activates a trace of all the IDMSQSAM look-ahead I/O reads. This trace shows the name of the file(s) being accessed by IDMSQSAM, each RBN that is read using QSAM or BDAM (DAM/EXCP), and a summary of the number of RBN's read through QSAM and BDAM. It also shows the area being accessed and the number of OPSYS QSAM buffers being used as determined by the JCL.

Default: OFF

REREAD_SYSCTL=ON|OFF

ON directs local mode operations to reread the SYSCTL file for each new transaction. This allows you to do the following:

  1. Include a SYSCTL in a batch job step's JCL.
  2. Start a transaction that will execute under central version, based on the contents of the SYSCTL file.
  3. Deallocate the SYSCTL file defined in the JCL.
  4. Start another transaction to execute in local mode.

Default: OFF

ROLLBACK3490

Enables the ROLLBACK utility to process archive files residing on devices that do not support backward read, such as disk and 3490E devices.

SAVE_SQL_SYNTAX=ON|OFF

If ON is specified during CV startup, all dynamic SQL syntax is saved before it gets optimized. To do so, the SQL server module IDMSQSRV allocates a piece of long term storage each time an sLTE (that is a secondary LTE representing the SQL session) is acquired. That storage is anchored off the sLTE and is used to save the dynamic SQL syntax before it is processed by the syntax parser and optimized.

The storage is obtained only once per session and is re-used if there is more SQL syntax to save for the same session. The storage is released during the end of session processing just prior to freeing the sLTE.

You can obtain the saved SQL syntax by Exit 39 - SQL Syntax Collecting Exit.

Default: OFF

SCRATCH_IN_STORAGE=ON|OFF|ANY|XA|64-bit

Enables storage allocation from the operating system for scratch processing.

ON

Specifies the same as SCRATCH_IN_STORAGE=ANY.

OFF

Specifies to allocate the scratch area as defined in the DMCL.

ANY

Acquires 64-bit storage if possible. If the request to allocate 64-bit storage fails, XA storage is acquired.

XA

Acquires 31-bit storage.

64-bit

Acquires 64-bit storage. If the request to allocate 64-bit storage fails, no attempt to acquire XA storage is done.

Default: OFF

Note: Usage of 64-bit storage is controlled by the MEMLIMIT parameter of the JOB or EXEC JCL statement on z/OS, and by the MEMLIMIT option on the SYSDEF MEMOBJ command on z/VSE.

SCRATCH_LIMIT=limit-with-unit

Specifies the maximum amount of scratch storage. The system continues to allocate more storage for scratch processing until the sum of all allocations reaches the value specified by limit-with-unit. Enter a number in the range 1-32767 followed by a unit of KB (Kilobyte: 2**10), MB (Megabyte: 2**20), GB (Gigabyte: 2**30), TB (Terabyte: 2**40), or PB (Petabyte: 2**50).

The default value is determined as follows:

SCRATCH_PRIMARY_EXTENT=prim-size-with-unit

Specifies the primary scratch allocation size. Enter a number in the range 1-32767 followed by a unit of KB (Kilobyte: 2**10), MB (Megabyte: 2**20), GB (Gigabyte: 2**30), TB (Terabyte: 2**40), or PB (Petabyte: 2**50).

The default value is determined as follows:

SCRATCH_SECONDARY_EXTENT=sec-size-with-unit

Specifies the amount of storage to allocate when all existing storage is in use. Enter a number in the range 1-32767 followed by a unit of KB (Kilobyte: 2**10), MB (Megabyte: 2**20), GB (Gigabyte: 2**30), TB (Terabyte: 2**40), or PB (Petabyte: 2**50).

The default size of the secondary allocation is equal to the size of the primary allocation.

SHUTDOWN_STALL_TIME=nnn

Causes the stall interval for active user tasks to be set to nnn seconds after a SHUTDOWN command is issued. If the parameter is not coded or an interval of 0 (zero) is specified, then there is no effect on SHUTDOWN processing. This parameter allows normal SHUTDONW processing even if a user task is in a long wait for an unavailable resource. The specified interval is used to force user tasks to be aborted if a given wait exceeds the interval even if the task normally has a longer STALL interval. This parameter affects online application tasks and ERUS tasks in either an INTERNAL or EXTERNAL wait. It does not have any effect on system tasks such as service drivers, line drivers, factotums, and helots.

SORT FIELD MAX LEN=n

n specifies the maximum sort key field length allowed by a sort.

Default (z/OS): 2000

Default (z/VSE): 256

SORTSIZE=ON|OFF

Directs whether CA IDMS utilities generate the SIZE= sort parameter card. Some sort packages cannot handle the SIZE= parameter. The default is OFF, which means that the SIZE= sort parameter is not generated.

Default: OFF

SQL_CACHE_ENTRIES=n

n specifies the maximum number of entries that will be used in the dynamic SQL cache. One entry holds one cached SQL statement. Specify 0 to disable caching. The maximum theoretical value is 2,147,483,647. The actual maximum depends on available memory.

Default: 200

SQL_INTLSORT=ON|OFF

Allows you to force the internal CA IDMS sort to be used in local mode. If ON is specified, an internal SORT rather than an operating system SORT will be performed on SQL commands issued in a local batch job that contain an ORDER BY clause. In many cases, an internal SORT is faster than an operating system SORT when you are not dealing with a large amount of data. OFF is the default, indicating an operating system SORT will be used.

Default: OFF

SQLTRACE=ON|OFF

ON activates a trace facility of all the SQL database requests made by an application.

Default: OFF

SYS_MSG=UPLOW|UPPER

UPPER directs CA IDMS to translate the text of internal #WTL messages to uppercase before being displayed at the output destination. The default is UPLOW. This allows the text of an internal #WTL message issued by CA software to be displayed in mixed case letters.

Specify UPPER under the following conditions:

SYSCTL=ddname

Specifies an alternate ddname for the SYSCTL file (other than the default ddname of SYSCTL).

SYSTRACE=ON|OFF

Controls whether basic system tracing is enabled.

ON

Enables basic system tracing.

Note: If basic system tracing is enabled by the SYSIDMS SYSTRACE parameter, it remains enabled for a system even if its system definition indicates that SYSTRACE is OFF.

OFF

Does not enable basic system tracing.

SYSTRK_DDNAME_PREFIX=xxxxxxx

Specifies the DDname prefix to be used for referencing SYSTRK files in execution JCL.

xxxxxxx

Specifies the 1- to 7-character DDname prefix.

Default: SYSTRK

TCP/IP_STACK_1=tcp/ip-stack-name
TCP/IP_STACK_2=tcp/ip-stack-name
TCP/IP_STACK_3=tcp/ip-stack-name
TCP/IP_STACK_4=tcp/ip-stack-name
TCP/IP_STACK_5=tcp/ip-stack-name
TCP/IP_STACK_6=tcp/ip-stack-name
TCP/IP_STACK_7=tcp/ip-stack-name
TCP/IP_STACK_8=tcp/ip-stack-name

In r16, the default TCP/IP stack name was extracted from the SYSTCPD file, and these 8 SYSIDMS parameters could be used to include 8 additional stacks to the CA IDMS TCP/IP environment. In r17, these parameters are still supported for compatibility reasons, but the method to define the TCP/IP stacks using the INCLUDE STACK clause in SYSGEN, or the two new SYSIDMS parameters INCLUDE_TCP/IP_STACK and EXCLUDE_TCP/IP_STACK will take precedence on these r16 definitions.

TCP/IP_STATUS=ON|OFF

Specifies the default status of TCP/IP support at startup. It overwrites the DEFAULT STATUS value defined to SYSGEN. OFF disables TCP/IP support for CA IDMS at startup. ON enables TCP/IP support.

TRACE_TABLE_SIZE=table-size KB | MB

Specifies the size of the system trace table in kilobytes (KB) or megabytes (MB).

Limit: 0–9999

Default: 0

Note: If basic system tracing is enabled by the SYSIDMS SYSTRACE parameter and the table size is 0, the table size is changed to 4 MB.

Any non-zero table size established by a SYSIDMS parameter overrides the trace table size specified in the system definition.

TRANSACTION_SHARING=ON|OFF

ON activates the 'Transaction Sharing' feature for all database activity used in a batch application.

Default: OFF

UPPER=INPUT|OUPUT|BOTH|OFF

Specifies whether input and/or output files will be converted to uppercase:

INPUT

Converts SYSIPT input files to uppercase.

OUTPUT

Converts SYSLST output files to uppercase.

BOTH

Converts both SYSIPT input files and SYSLST output files to uppercase.

OFF

Does not convert SYSIPT input files or SYSLST output files to uppercase.

Default: OFF

USERCAT=ON|OFF

Specifies whether the user catalog is to be accessed. Specify OFF only when formatting the user catalog or when the DMCL does not have access to a user catalog.

Default: ON

WIDTH_PAGE=nnn

Specifies a maximum number of characters to be printed on a SYSLST output line.

Limits: 71-132

Default: 132

WORK=n

CA IDMS utilities that generate SORT input cards for use within a VSE environment typically hardcode the number of SORTWK files to be allocated. This value can be insufficient for the amount of data to be sorted. This parameter allows sites to specify the number of SORTWK files to be allocated by their SORT.

Limits: 1-9

Note: Any other values are ignored and the default value of the utility is used instead.

XA_SCRATCH=ON|OFF

Specifies whether scratch space will be allocated out of XA storage or not. OFF, the default, indicates that a scratch file will be used.

Default: OFF

Note: XA_SCRATCH=ON/OFF is maintained for compatibility reasons only. Use SCRATCH_IN_STORAGE instead.

BUFFERSTAT Report Field Descriptions

The following table gives the descriptions for the fields displayed on the report produced by the BUFFERSTAT SYSIDMS parameter.

*** Buffer Name ***

The name of the buffer from the DMCL which has been opened during the processing of this job. Only those buffers which are open or have had some I/O activity against them will appear in this report.

*** Pages ***

The number of pages allocated to the buffer. This is the actual number of pages 'in use' by this buffer for this job. This number is the total from the a) DMCL Local Mode Buffer Pages nnn, or b) JCL DCB=BUFNO=nnn, or c) SYSIDMS Parm FILE_BUF=ddname=nnn.

*** Prefetch Minimum ***

The minimum number of buffers 'in use' needed to allow Prefetch to operate from Random access verbs (that is, non-area sweep processing).

DB Page Requests

The number of times IDMSDBMS requests a database page from the buffer pool by calling IDMSDBIO.

Sequential Area Request

Of the DB Page requests in the 'DB Page Requests' count, how many were GET/NEXT/PRIOR in AREA verbs.

Random Request

Of the DB Page requests in the 'DB Page Requests' count, how many were not GET/NEXT/PRIOR in AREA verbs.

Found in Buffer

Of the DB Page requests in the 'DB Page Requests' count, how many DB pages were already in the buffer pool and did not require an I/O.

Not Found in Buffer

Of the DB Page requests in the 'DB Page Requests' count, how many DB pages were not in the buffer pool and therefore required an I/O.

Buffer 'hit' Ratio

Calculated as 'Found in Buffer' times 100, divided by the DB Page Requests value.

Found in Pref Buffer

Found in Prefetch Buffer. Of the 'DB Page Requests' count, how many DB pages were found in the buffer pool that had been read by a Prefetch Read?

Found in Cache

Of the 'DB Page Requests' count, how many DB pages were found in the shared cache?

Total DB Pages Read

The total number of DB pages read by both Prefetch and standard I/O. This is not the number of I/O's or EXCP's, but the number of DB pages read into a buffer as a standalone I/O, 'Start I/O - Reads', or as part of a Prefetch I/O 'Pages Read - Prefetch'.

DB Pg Req:Tot Pages Read

The number of 'DB Page Requests' count divided by the 'Total DB Pages Read' count.

DB Pg Req:Strt I/O Read

The number of 'DB Page Requests' count divided by the 'Start I/O Reads' count.

NonPrefetch I/O Rqst

The number of standard of non-Prefetch I/O requests. This is the number of the 'DB Pages Request' count which are 'Not found in Buffer' and were not eligible for Prefetch. Prefetch was either not allowed, turned off, or the request was started by a 'Random Request' for which the '*** Prefetch Minimum ***' is higher than the number of '*** Pages ***'.

Start I/O - Reads

The number of standard or non-Prefetch I/O reads. This number is the result of 'DB Page Requests' which were 'Not found in Buffer' and Prefetch is turned off, or the request was started by a 'Random Request' for which the '*** Prefetch Minimum ***' is higher than the number of '*** Pages ***' count, or the request is not eligible for Prefetch processing. Each of these will show up as 1 EXCP.

Start I/O - Writes

The number of writes started against the database. Each of these will show up as 1 EXCP.

Read forces Write

In order to read a DB page into the buffer, a DB page had to be written out to the database first (based on the least recently used algorithm). When this occurs, Prefetch is effectively turned off.

Prefetch Requests

With Prefetch operating, the number of DB page requests 'Not found in Buffer', which were eligible for Prefetch processing.

Sequential Area Request

Of the 'Prefetch Requests' count, how many were GET NEXT/PRIOR in AREA type verbs.

Random Request

Of the 'Prefetch Requests' count, how many were not GET NEXT/PRIOR in AREA type verbs.

Pref Req Denied:Buffers

Of the 'Prefetch Requests' count, how many did not use Prefetch due to too many buffer pages with the 'must write switch' on (over 1/2 the number of pages in the buffer pool).

Start I/O - Prefetch

The number of the 'Prefetch Requests' count that actually "Start an I/O" or "Start Subchannel". Each of these will show up as 1 EXCP.

Pages Read - Prefetch

The number of DB pages that were "carried" with every 'Start I/O - Prefetch'. This number plus the 'Start I/O - Reads' will equal the 'Total DB Pages Read' count.

Pref Strt I/O:Pref Req

The 'Start I/O - Prefetch' count divided by the 'Prefetch Requests' count.

Pref Pages:Pref Strt I/O

The 'Pages Read - Prefetch' count divided by the 'Prefetch Requests' count.

Pref Pages:Pref Strt I/O

The 'Pages Read - Prefetch' count divided by the 'Start I/O - Prefetch' count. This value shows how effective the Prefetch operation is. Compare this number to the "pages per track" to see how effective each Prefetch I/O is. If this number is around 3/1 or less, you probably will not see enough improvement in performance to warrant using Prefetch.

For each file in use, there will be a set of counts:

z/VSE File Parameters

The following parameters can be used to override default values for z/VSE work files and for the system logical units SYSRDR, SYSIPT, SYSPCH, and SYSLST in batch jobs only.

FILENAME=file-name

Specifies the name of the file whose attributes are to be overridden by the following SYSIDMS parameters.

BLKSIZE=block-size

Specifies the block size for a file. BLKSIZE and BLOCKS are mutually exclusive parameters.

BLOCKS=block-count

Specifies a blocking factor for a file. BLKSIZE and BLOCKS are mutually exclusive parameters.

DEVADDR=SYSxxx

Specifies a device address for a tape file (SYSIPT, SYSLST, SYSRDR, SYSPCH, or SYSlogical-unit-number).

FILABL=STD/NO

Specifies a no-label option for a tape file. FILABL=STD is the default.

Default: FILABL=STD

FILETYPE=T/D/I

Specifies a file type of tape, disk, or file independent.

LRECL=logical-record-size

Specifies the logical record size for a file.

RECFM=F/V

Specifies if the current file contains fixed-length or variable-length records. V is the default.

Default: V

REWIND=YES/NO/UNLOAD

Specifies the position of a tape file when it is opened or closed. REWIND=UNLOAD is the default.

Default: REWIND=UNLOAD