Use CSQ6SYSP
Use this topic as a reference for how to set system parameters using CSQ6SYSP.
The default parameters for CSQ6SYSP, and whether we can alter each parameter using the SET SYSTEM command, are shown in Table 1. To change any of these values, see the detailed descriptions of the parameters.Parameter | Description | Default value | SET command |
---|---|---|---|
ACELIM | Size of ACE storage pool in 1 KB blocks. | 0 (no limit) | |
CLCACHE | Specifies the type of cluster cache to use. | STATIC | - |
CMDUSER | The default user ID for command security checks. | CSQOPR | - |
EXCLMSG | Specifies a list of messages to be excluded from any log. Messages in this list are not sent to the z/OS console and hardcopy log. As a result using the EXCLMSG parameter to exclude messages is more efficient from a CPU perspective than using the methods described in Suppress information messages | ( ) | |
EXITLIM | Time (in seconds) for which queue manager exits can run during each invocation. | 30 | - |
EXITTCB | How many started server tasks to use to run queue manager exits. | 8 | - |
LOGLOAD | Number of log records written by IBM MQ between the start of one checkpoint and the next. | 500 000 | |
MULCCAPT | Determines the Measured Usage Pricing property which controls the algorithm for gathering data used by Measured Usage License Charging (MULC). | See parameter description | - |
OTMACON | OTMA connection parameters. | See parameter description | - |
QINDXBLD | Determines whether queue manager restart waits until all indexes are rebuilt, or completes before all indexes are rebuilt. | WAIT | - |
QMCCSID | Coded character set identifier for the queue manager. | Zero | - |
QSGDATA | Queue sharing group parameters. | See parameter description | - |
RESAUDIT | RESLEVEL auditing parameter. | YES | - |
ROUTCDE | Message routing code assigned to messages not solicited from a specific console. | 1 | - |
SERVICE | Reserved for use by IBM. | 0 | |
SMFACCT | Specifies whether SMF accounting data is to be collected when the queue
manager is started.
Note that class 4 channel accounting data is collected only when the channel initiator is started. |
NO | - |
SMFSTAT | Specifies whether SMF statistics are to be collected when the queue manager is
started.
Note that class 4 channel initiator statistics data is collected only when the channel initiator is started. |
NO | - |
SPLCAP | Specifies whether queue security policy capability is enabled on this queue manager. For Advanced Message Security for z/OS, set this parameter to YES. | NO | - |
STATIME | Default time, in minutes, between each gathering of statistics. | 30 | |
TRACSTR | Specifies whether tracing is to be started automatically. | NO | - |
TRACTBL | Size of trace table, in 4 KB blocks, to be used by the global trace facility. | 99 (396 KB) | |
WLMTIME | Time between scanning the queue index for WLM-managed queues. | 30 | - |
WLMTIMU | Units (minutes or seconds) for WLMTIME. | MINS | - |
- ACELIM
- Specifies the maximum size of the ACE storage pool in 1 KB blocks. The number must be in the
range 0-999999. The default value of zero means no imposed constraint, beyond what is available in
the system.
We should only set a value for ACELIM on queue managers that have been identified as using exorbitant quantities of ECSA storage. Limiting the ACE storage pool has the effect of limiting the number of connections in the system, and so, the amount of ECSA storage used by a queue manager.
Once the queue manager reaches the limit it is not possible for applications to obtain new connections. The lack of new connections causes failures in MQCONN processing, and applications coordinated through RRS are likely to experience failures in any IBM MQ API.
An ACE represents approximately 8% of the total ECSA required for the thread-related control blocks for a connection. So, for example, specifying ACELIM=5120 would be expected to cap the total amount of ECSA allocated by the queue manager (for thread-related control blocks) at approximately 64000K; that is 5120 multiplied by 12.5.
In order to cap the amount total amount of ECSA allocated by the queue manager, for thread-related control blocks at 5120K, an ACELIM value of 410 is required.
We can use SMF 115 subtype 5 records, produced by statistics CLASS(3) trace, to monitor the size of the 'ACE/PEB' storage pool, and hence set an appropriate value for ACELIM.
We can obtain the total amount of ECSA storage used by the queue manager, for control blocks, from SMF 115 subtype 7 records, written by statistics CLASS(2) trace. The total amount of ECSA storage used is the sum of the QSRSPHBGF and QSRSPHBGV fields.
For more information about SMF 115 statistics records, see Interpreting IBM MQ performance statistics.
Note that, we should consider setting ACELIM as a mechanism to protect a z/OS image from a badly behaving queue manager, rather than as a means to control application connections to a queue manager.
- CLCACHE
- Specifies the type of cluster cache to use. See Configure a queue manager cluster for more
information.
- STATIC
- When the cluster cache is static, its size is fixed at queue manager start-up, enough for the current amount of cluster information plus some space for expansion. The size cannot increase while the queue manager is active. This is the default.
- DYNAMIC
- When the cluster cache is dynamic, the initial size allocated at queue manager startup can be increased automatically if required while the queue manager is active.
- CMDUSER
-
Specifies the default user ID used for
command security checks. This user ID must be defined to the ESM (for example, RACF ). Specify a name of 1 through 8 alphanumeric
characters. The first character must be alphabetic.
The default is CSQOPR.
- EXCLMSG
-
Specifies a list of error messages to be excluded.
This list is dynamic and is updated using the SET SYSTEM command.
The default value is an empty list ( ).
Messages are supplied without the CSQ prefix and without the action code suffix (I-D-E-A). For example, to exclude message CSQX500I, add X500 to this list. This list can contain a maximum of 16 message identifiers.
To be eligible to be included in the list, the message must be issued after normal startup of the MSTR or CHIN address spaces and begin with the one of the following characters E, H, I, J, L, M, N, P, R, T, V, W, X, Y, 2,3, 5, 9.
Message identifiers that are issued as a result of processing commands can be added to the list, however will not be excluded. For example, a message identifier is issued as a result of the DISPLAY USAGE PSID(*) command, however, this message can not be suppressed.
- EXITLIM
-
Specifies the
time, in seconds, allowed for each invocation of the queue manager exits. (This parameter has no
effect on channel exits.)
Specify a value in the range 5 through 9999.
The default is 30. The queue manager polls exits that are running every 30 seconds. On each poll, any that have been running for more than the time specified by EXITLIM are forcibly terminated.
- EXITTCB
-
Specifies the number of started server tasks to
use to run exits in the queue manager. (This parameter has no effect on channel exits.) We must
specify a number at least as high as the maximum number of exits (other than channel exits) that the
queue manager might have to run, else it will fail with a 6c6 abend.
Specify a value in the range zero through 99. A value of zero means that no exits can be run.
The default is 8.
- LOGLOAD
-
Specifies the number of log records
that IBM MQ writes between the start of one checkpoint
and the next. IBM MQ starts a new checkpoint after the
number of records that you specify has been written.
Specify a value in the range 200 through 16 000 000.
The default is 500 000.
The greater the value, the better the performance of IBM MQ ; however, restart takes longer if the parameter is set to a large value.
Suggested settings:Test system 10 000 Production system 500 000 In a production system, the supplied default value might result in a checkpoint frequency that is too high.
The value of LOGLOAD determines the frequency of queue manager checkpoints. Too large a value means that a large amount of data is written to the log between checkpoints, resulting in an increased queue manager forward recovery restart time following a failure. Too small a value causes checkpoints to occur too frequently during peak load, adversely affecting response times and processor usage.
An initial value of 500 000 is suggested for LOGLOAD. For a 1 KB persistent message rate of 100 messages a second (that is, 100 MQPUT s with commit and 100 MQGET s with commit) the interval between checkpoints is approximately 5 minutes.
Note: This is intended as a guideline only and the optimum value for this parameter is dependent on the characteristics of the individual system. - MULCCAPT
-
Specifies the algorithm to be used for
gathering data used by Measured Usage License Charging (MULC).
- STANDARD
- MULC is based on the time from the IBM MQ API MQCONN call to the time of the IBM MQ API MQDISC call.
- REFINED
- MULC is based on the time from the start of an IBM MQ API call to the end of the IBM MQ API call.
The default is STANDARD
- OTMACON
-
OTMA parameters. This
keyword takes five positional parameters::
- OTMACON = ( Group,Member,Druexit,Age,Tpipepfx)
-
- Group
-
This is the name
of the XCF group to which this particular instance of IBM MQ belongs.
It can be 1 through 8 characters long and must be entered in uppercase characters.
The default is blanks, which indicates that IBM MQ must not attempt to join an XCF group.
- Member
-
This is the
member name of this particular instance of IBM MQ within
the XCF group.
It can be 1 through 16 characters long and must be entered in uppercase characters.
The default is the 4-character queue manager name.
- Druexit
-
This specifies the name of
the OTMA destination resolution user exit to be run by IMS.
It can be 1 through 8 characters long.
The default is DFSYDRU0.
This parameter is optional; it is required if IBM MQ is to receive messages from an IMS application that was not started by IBM MQ. The name must correspond to the destination resolution user exit coded in the IMS system. For more information see Use OTMA exits in IMS.
- Age
-
This represents the length
of time, in seconds, that a user ID from IBM MQ is
considered previously verified by IMS.
It can be in the range zero through 2 147 483 647.
The default is 2 147 483 647.
You are recommended to set this parameter in conjunction with the interval parameter of the ALTER SECURITY command to maintain consistency of security cache settings across the mainframe.
- Tpipepfx
-
This represents the prefix to be
used for Tpipe names.
It comprises three characters; the first character is in the range A through Z, subsequent characters are A through Z or 0 through 9. The default is CSQ.
This is used each time IBM MQ creates a Tpipe; the rest of the name is assigned by IBM MQ. We cannot set the full Tpipe name for any Tpipe created by IBM MQ.
- QINDXBLD
-
Determines whether queue
manager restart waits until all queue indexes are rebuilt, or completes before all indexes are rebuilt.
- WAIT
-
Queue manager restart waits for
all queue index builds to be completed. This means that no applications are delayed during normal
IBM MQ API processing while the index is created, as all
indexes are created before any applications can connect to the queue manager.
This is the default.
- NOWAIT
- The queue manager can restart before all queue index building is completed.
- QMCCSID
-
Specifies the default coded character set identifier that the queue manager (and
therefore distributed queuing) is to use.
Specify a value in the range zero through 65535. The value must represent an EBCDIC code page listed as a native z/OS code page for the chosen language in National languages.
Zero, which is the default value, means use the CCSID currently set or, if none is set, use CCSID 500. This means that if you have explicitly set the CCSID to any non-zero value, we cannot reset it by setting QMCCSID to zero; we must now use the correct non-zero CCSID. If QMCCSID is zero, we can check what CCSID is actually in use by issuing the command DISPLAY QMGR CCSID.
- QSGDATA
-
Queue sharing
group data. This keyword takes five positional parameters:
- QSGDATA=( Qsgname,Dsgname,Db2name,Db2serv,Db2blob)
-
- Qsgname
-
This is the name of the
queue sharing group to which the queue manager belongs. See Rules for naming IBM MQ objects for valid characters. The name:
- Can be 1 through 4 characters long
- Must not start with a numeric
- Must not end in @.
This is because, for implementation reasons, names of less than four characters are padded internally with @ symbols,
The default is blanks, which indicates that the queue manager is not a member of any queue sharing group.
- Dsgname
-
This is
the name of the Db2 data-sharing group to which the
queue manager is to connect.
It can be 1 through 8 characters long and must be entered in uppercase characters.
The default is blanks, which indicates that we are not using queue sharing groups.
- Db2name
-
This is the name of the Db2 subsystem or group attachment to which the queue manager
is to connect.
It can be 1 through 4 characters long and must be entered in uppercase characters.
The default is blanks, which indicates that we are not using queue sharing groups.
Note: The Db2 subsystem (or group attachment) must be in the Db2 data-sharing group specified in the Dsgname, and all queue managers must specify the same Db2 data-sharing group. - Db2serv
-
This is the number of server
tasks used for accessing Db2.
It can be in the range 4 through 10.
The default is 4.
- Db2blob
-
This is the number of
Db2 tasks used for accessing Binary Large Objects
(BLOBs).
It can be in the range 4 through 10.
The default is 4.
If you specify one of the name parameters (that is, Qsgname, Dsgname, or Db2name ), we must enter values for the other names, otherwise IBM MQ fails.
- RESAUDIT
-
Specifies whether RACF audit records are written for RESLEVEL security checks
performed during connection processing. Specify one of:
- NO
- RESLEVEL auditing is not performed.
- YES
- RESLEVEL auditing is performed.
The default is YES.
- ROUTCDE
-
Specifies the default z/OS message routing code assigned to messages that are not
sent in direct response to an MQSC command. Specify one of:
- A value in the range 1 through 16, inclusive.
- A list of values, separated by a comma and enclosed in parentheses. Each value must be in the range 1 through 16, inclusive.
The default is 1.
For more information about z/OS routing codes, see the MVS Routing and Descriptor Codes manual.
- SERVICE
- This field is reserved for use by IBM.
- SMFACCT
-
Specifies whether
IBM MQ sends accounting data to SMF automatically when
the queue manager starts. Specify one of:
- NO
- Do not start gathering accounting data automatically.
- YES
- Start gathering accounting data automatically for the default class 1.
- integers
- A list of classes for which accounting is gathered automatically in the range 1 through 4.
The default is NO.
- SMFSTAT
-
Specifies whether to
gather SMF statistics automatically when the queue manager starts. Specify one of:
- NO
- Do not start gathering statistics automatically.
- YES
- Start gathering statistics automatically for the default class 1.
- integers
- A list of classes for which statistics are gathered automatically in the range 1 through 4. To gather class 2 or 3 statistics, class 1 must also be specified.
The default is NO.
- SPLCAP
-
The security policy capability enables higher
level of message security through policies that control whether messages are signed, or encrypted,
as they are written and read from queues. Security policy processing is enabled for this queue
manager, by configuring SPLCAP with one of the following values:
- NO
- The capability to implement message security policies for queues is not enabled during queue manager initialization.
- YES
- Message security capabilities are enabled during queue manager initialization.
If this control is set, the queue manager attempts to load the license enabling module from SDRQAUTH during initialization, and start an additional address space (AMSM).
The queue manager does not start unless AMS is licensed, and the necessary configuration for message security is in place.
The default is NO.
- STATIME
-
Specifies the
default time, in minutes, between consecutive gatherings of statistics.
Specify a number in the range zero through 1440.
If you specify a value of zero, both statistics data and accounting data is collected at the SMF data collection broadcast. See Use System Management Facility for information about setting this.
The default is 30.
- TRACSTR
-
Specifies whether
global tracing is to start automatically. Specify one of:
- NO
- Do not start global tracing automatically.
- YES
- Start global tracing automatically for the default class, class 1.
- integers
- A list of classes for which global tracing is to be started automatically in the range 1 through 4.
- *
- Start global trace automatically for all classes.
The default is NO if we do not specify the keyword in the macro.
Note: The supplied default system parameter load module (CSQZPARM) has TRACSTR=YES (set in the assembler module CSQFSYSP). If we do not want to start tracing automatically, either create your own system parameter module, or issue the STOP TRACE command after the queue manager has started.For details about the STOP TRACE command, see STOP TRACE.
- TRACTBL
-
Specifies the default size, in 4 KB blocks,
of trace table where the global trace facility stores IBM MQ trace records.
Specify a value in the range 1 through 999.
The default is 99. This is equivalent to 396 KB.
Note: Storage for the trace table is allocated in the ECSA. Therefore, we must select this value with care. - WLMTIME
-
Specifies the time (in minutes or
seconds depending on the value of WLMTIMU) between each scan of the indexes for WLM-managed queues.
Specify a value in the range 1 through 9999.
The default is 30.
- WLMTIMU
-
Time units used with the WLMTIME
parameter. Specify one of :
- MINS
- WLMTIME represents a number of minutes.
- SECS
- WLMTIME represents a number of seconds.
The default is MINS.
Parent topic: Tailor the system parameter module
Related reference