Application Performance Analyzer for z/OS, Version 11.1, Customization Guide

Started task configuration settings

The Application Performance Analyzer started task CAZ0 obtains these settings from a file allocated to DDname CONFIG. The file is a SYSIN-type member which you edit in order to change the settings. The member CAZCNFG0 in hlq.SCAZSAMP contains the initial settings. A group of configuration settings is specified by a CONFIG statement. Settings are organized into categories or classes. Each CONFIG statement will define settings for a particular class.

The general syntax for a CONFIG statement is illustrated here:

CONFIG class
keyword1=value1
keyword2=value2    // this is a comment
..
keywordn=valuen ;  // end with semicolon

You must code a single keyword/value pair expression within a single line; it cannot span two lines. Separate keyword/value pair expressions from each other using one or more blanks. A CONFIG statement, including all its keyword/value pair expressions must be terminated by a semicolon.

A CONFIG statement might be coded anywhere within columns 1 to 72 inclusive. An asterisk in column 1 denotes a comment line. A double slash (//) anywhere in the line indicates the remaining text in the line is a comment.

The following CONFIG BASIC statement, which is supplied in the predefined CAZCNFG0 member, illustrates the CONFIG syntax.

CONFIG BASIC                 // Basic settings:
   STCid=CAZ0                // . Application Performance Analyzer instance ID
   HLQ=CAZ0                  // . DSN high level qualifier
   UNIT=SYSDA                // . unit name for new data sets
   CheckpointDSN='.CHKP';    // . checkpoint file DSN
                             // (appended to HLQ)

CONFIG statement parameters

Class
Specifies the category of settings that the CONFIG statement is to modify. The allowable values for class are:
keywordn
This specifies the name of the setting variable and is paired with a value expression in keyword=value format. The name is not case sensitive.
valuen
This is paired with a keyword name and specifies the value to be set.

CONFIG BASIC statement

BASIC settings allow you specify values for basic characteristics of the Application Performance Analyzer installation. The following settings can be specified:

Note:
Required parameters are listed first. The alphabetical order specified below is not the order of the parameters in the configuration file.
CDS
Required.

This setting indicates whether the common data store is used for source program mapping, and if so specifies the name given to the Application Performance Analyzer common data store file.

The syntax for the CDS keyword is CDS=(status,dsn) where

For further information on the common data store for source program mapping, refer to Chapter 10 in the Application Performance Analyzer for z/OS User’s Guide.

The following example enables the common data store with the file name of IBMAPA.CDS:

CDS=(ENA,'.CDS')
HLQ=IBMAPA

The following example disables the common data store:

CDS=(DIS)

Default: none

CheckpointDSN
Required.

Specifies the name to be given to the Application Performance Analyzer checkpoint file. You have two options when specifying this value:

Default: None

HLQ
Required.

Specifies the high-level qualifier for data sets created by the Application Performance Analyzer started task as part of its normal operation. The value that you specify for HLQ must comply with the JCL rule for specifying a data set name. It can be no more than 17 characters long.

Note:
If this Application Performance Analyzer instance is a member of a sysplex group, you must select a high-level qualifier that will allow Application Performance Analyzer to allocate all new data sets to volumes that are shared across the sysplex. These data sets include the checkpoint file, sample files, and optional log files. See SampleHLQ in CONFIG SAMPLE statement, and LogHLQ in CONFIG LOGFILE statement for descriptions of how to set a unique high-level qualifier for the sample files and log files.

Default: None.

STCID
Required.

This is the Application Performance Analyzer ID, a four-character value that uniquely identifies the Application Performance Analyzer started task.

Note:
If you intend to run Application Performance Analyzer in a SYSPLEX group, you must never change the value of STCID once it has been set. The STCID forms part of the Application Performance Analyzer SYSPLEX group name. Application Performance Analyzer compares the group name recorded in each observation record, with the active Application Performance Analyzer group name to determine if the observation is local or remote to that instance of Application Performance Analyzer. Changing the STCID will cause observations once local to a Application Performance Analyzer instance to be treated as remote. Under rare conditions, this may potentially trigger a high volume of unnecessary XCF messages.
DeleteOnJCLError
Optional.

Specifies whether multiple-step requests that get a JCL error in one of the steps should delete all the sample files and entries for the remaining steps. Specify DeleteOnJCLError=Y to have the entries for steps with a JCL error automatically deleted.

Default: DeleteOnJCLError=N

ExpiryDays
Optional.

This setting indicates how Application Performance Analyzer is to manage the retention of completed observations. It accepts 3 variables using the syntax: ExpiryDays=(days,ExpiryProt,WARN=nnn) where

days = 0-9999
This specifies the number of days an observation is to be retained, before it is automatically deleted by Application Performance Analyzer. The user can override this value when entering an observation request by overtyping the Retain File for (days) field on the Job Information NEW panel. Specify 0 to retain the observation indefinitely. 0 is the default.
ExpiryProt = Y|N
This specifies whether the Retain File for (days) field on the Job Information NEW panel is protected or unprotected. Specify Y to protect the field from changes. Specify N to leave the field unprotected. N is the default.
nnn = 0-255.
This enables the expiry days warning feature and specifies the number of days prior to observation automatic deletion that Application Performance Analyzer is to start warning the user. When this feature is enabled, the request number in the observation list might be displayed in red, yellow or green based on the expiry date. It is displayed in red when an observation will be automatically deleted within 24 hours. It will be displayed in yellow when an observation will be automatically deleted within the number of days specified for WARN. It will be displayed in green when the observation will never be deleted. Users with administrator authority will see the colors for all users' observations. All other users will see the colors for their own observations only. A value of 255 disables this feature. 255 is the default.

Examples:

Default: ExpiryDays=(0,N,WARN=255) to indicate the observation file is to be retained indefinitely, the Retain File for (days) field is unprotected, and the expiry day warning feature is disabled.

MsgLevel
Optional.

Specifies the severity level of messages to be written to the log data sets. MsgLevel=I will write all messages, MsgLevel=W will write warning level and higher severity messages only. No SYSOUT data sets will be created for a sampling session with MsgLevel=W when only informational messages are produced.

Default: MsgLevel=I

SecureImportSys
Optional.

Specify Y (yes) to use the originating system name of imported sample files in Application Performance Analyzer security requests. Specify N (no) to use the local system name, which is the system into which the sample file is imported. This applies only when Security=External is specified. When Y, external security rules must be written to accommodate the use of the originating system name.

This keyword is applicable in cases where sample files are imported from an originating system with restricted access into a local system that has less restrictive access. If SecureImportSys=Y is not specified, users that are not authorized to view sample files on the originating system could potentially view the same sample file once imported into the local system.

Default: N

Example: System PROD allows only users USERA and USERB to view Application Performance Analyzer reports. System TEST allows all users to view all Application Performance Analyzer reports. To ensure that only USERA and USERB can view reports in TEST for sample files imported from PROD to TEST, the following must be specified in the TEST system:

Security
Optional. (But Security=External is highly recommended.)

Specifies whether this started task should use Application Performance Analyzer's internal security scheme or use an external product, such as RACF. Specify Security=External if you use an external security product. To use Application Performance Analyzer's internal security scheme, you can specify Security=Internal or allow the default. (Application Performance Analyzer defaults to internal security.) You can also specify Security=None to run without security.

Default: Internal.

Note:
The Security=Internal option is supported for backwards compatibility, but is not recommended. Similarly, you can specify the keyword COMPAT along with Security=External for backwards compatibility with an early version of the external security feature. The list below explains the available security options:
Security=Internal
Specifies that no external security will be used. This option is not recommended, particularly in sysplex environments.
Security=External
Specifies that external security will be used. This is the recommended option. It fully supports sysplex environments.
Security=(External,Compat)
Specifies that external security will be used. This option is for Application Performance Analyzer, Version 1.1 with PTF UK07430 applied for APAR PK10866 only. Do not use this option if you have not applied PTF UK07430, or if you have applied subsequent PTFs. (For external security with subsequent PTFs, use Security=External.) In this mode, the first node of a resource name is the STCID instead of the system name. For improved performance, migrate to the Security=External option.
Security=None
Specifies that no security will be used. This option is useful in test environments, in which security is not required.
Default: Internal
SecurityClassName
Optional.

Specifies an alternate class name will be used instead of FACILITY. This keyword is only used when Security=External is also specified; otherwise it is ignored. This is mandatory when SecurityHLQ is specified.

Default: FACILITY

Rules for class name:

  1. Must be between 1-8 characters, consisting of:
    • A-Z
    • 0-9
    • #, @, $
  2. Must contain at least one character from the following:
    • 0-9, #, @, $

If SecurityClassName is coded as FACILITY or allowed to default, then parameter SecurityHLQ may not be used. The FACILITY class is limited to 39 characters for the access rule that can be exceeded if SecurityHLQ is used. An alternative class is XFACILIT, a predefined class in RACF that allows up to 246 characters.

SecurityHLQ
Optional.

Specifies an optional qualifier for security profiles to be placed in front of the sysplex system name when a SecurityClassName has been specified other than FACILITY. This can be a maximum of 17 characters and must follow the rules established for the class in the Class Descriptor Table (CDT). The contents of this field are not checked during startup. This HLQ can be used in a sysplex environment to provide common access rules for all images in the sysplex.

If specified, it will be concatenated in front of the sysplex system name during security checks.

When SecuirtyHLQ is specified, SecurityClassName must also be specified, and it must not be FACILITY. See SecurityClassName for more details.

Default: None.

SecuritySuppMsg
Optional.

This option causes Application Performance Analyzer to either suppress or display messages issued as a result of RACROUTE calls made by Application Performance Analyzer. If not suppressed, messages will be displayed by RACROUTE via WTO.

Default: Y to suppress these messages.

SPXGroupName
Optional.

Specifies the Application Performance Analyzer sysplex group that this Application Performance Analyzer instance will be a member of. It can be a maximum of eight alphanumeric characters. All Application Performance Analyzer instances that will communicate with each other must code an identical SPXGroupName=sysplexname.

Note:
The value specified for SPXGroupName must be unique and not currently used by any other subsystems in the sysplex. Do not use the system's sysplex name for this value.

The Application Performance Analyzer instances in an Application Performance Analyzer sysplex group must be from different MVS™ systems. All Application Performance Analyzer instances defined in the same Application Performance Analyzer sysplex group must share a common checkpoint file; therefore, it is important that the HLQ, UNIT, and CheckpointDSN are identical.

In summary, the CONFIG BASIC statement for every Application Performance Analyzer instance defined in the same Application Performance Analyzer sysplex group must be coded identically, with the exception of the STCID, which might be unique.

Default: None.

TZ
Optional (only required for USS source mapping).

Specifies the TZ environment variable, which conforms to the POSIX standard. This must be coded if USS programs are to be source mapped using debug files in ELF/DWARF format. Otherwise, it is not required. Details on the format and content of a TZ environment variable can be found at:

 http://www.opengroup.org/onlinepubs/009695399/basedefs/xbd_chap08.html

.

TZ="EST5EDT,M3.2.0/02:00,M11.1.0/02:00"
UNIT
Optional.

Specifies the unit name to be used when allocating all new data sets of the following type:

Default: SYSDA

Unit can be either a device type such as 3390, or an esoteric such as SYSDA. A mount status of STORAGE or PUBLIC is nonspecific and does not require a VOLSER. (This release of Application Performance Analyzer does not allow a VOLSER to be specified.)

On a non-SMS-managed system, you must ensure that the UNIT name you specified contains one or more volumes mounted as either STORAGE or PUBLIC. If there are not such volumes mounted, then Application Performance Analyzer will be unable to allocate critical files and will terminate.

On an SMS-managed system, SMS will ignore the unit parameter. To ensure SMS control over all data sets allocated by Application Performance Analyzer, specify an HLQ value that is recognized and managed by SMS.

CONFIG SAMPLE statement

These settings pertain to the Sampling (measurement) process. The following settings might be specified:

Note:
The alphabetical order specified below is not the order of the parameters in the configuration file.
Adabas
Optional.

Specify YES to make the Adabas data extractor available.

Default: NO

AllExtractorsOn
Optional.

Specify YES to force all new measurement requests to have all the data extractors selected by default. This parameter should be used in special cases in which a particular instance of Application Performance Analyzer will be used to measure an address space by someone who does not know what activity types are running in that address space. Using AllExtractorsOn ensures that all types of activity will be recorded without the user manually selecting all the data extractors.

Default: NO

AltCICSStepNm
Optional.

Specifies an alternate load module name for the CICS region. By default, a step name of DFHSIP is used to determine if the region being sampled is a CICS region. If this step name is not found, CICS sampling is turned off to reduce overhead.

Default: DFHSIP

BufSize
Optional.

Specifies the size that Application Performance Analyzer is to use as the blocksize when allocating a new sample file.

Default: 16,384 bytes.

CICSI
Optional.

This is used to specify that the CICS Intercept should be turned on. Specify CICSI=YES to turn on this feature and to make the CICS+ data extractor available. When this feature is turned on and the CICS+ data extractor is selected for a measurement, Application Performance Analyzer will count the precise number of CICS transactions executed during a measurement and record the exact service time and CPU time for each transaction.

Default: NO

Compress
Optional.

Specify YES to turn sample file compression on. When sample file compression is activated, Application Performance Analyzer will compress all sample files prior to writing them. When a sample file is opened for reporting, Application Performance Analyzer will automatically decompress the file for the duration of the reporting session.

Default: NO

DB2EXPLAIN
Optional. Multiple DB2EXPLAIN entries are allowed.

The DB2EXPLAIN parameter must be specified to use the DB2® EXPLAIN feature in DB2 reports. (DB2+ must also be specified.)

The syntax for the DB2EXPLAIN keyword is as follows:

DB2EXPLAIN=(ssid,D|Q|A,Yes|No,Yes|No,dbname,tsname)

This entry is made up of from one to six positional parameters. The first parameter specifies a DB2 subsystem name. The next five positional parameters are optional.

The second parameter indicates the SQL STRING DELIMITER value specified in the application programming defaults panel (DSNTIPF) for the specified DB2 subsystem at installation. The values are as follows:

The default value is D. This value must match the DB2 subsystem installation value.

The third parameter indicates the USE FOR DYNAMICRULES value specified in the application programming defaults panel (DSNTIPF) for the specified DB2 subsystem at installation. The values are as follows:

The default value is Yes. This value must match the DB2 subsystem installation value.

The fourth parameter specifies whether automatic PLAN_TABLE creation is required for the DB2 subsystem. This does not apply to DB2, Version 8, except when it is in COMPAT mode. A value of Y or Yes specifies that dynamic creation is required. The parameter can be omitted if dynamic creation is not required.

The fifth parameter specifies the database name to be used for automatic PLAN_TABLE creation. It only applies if the fourth parameter has a value of Y or Yes and there is a database name specified in the fifth parameter.

The sixth parameter specifies the tablespace name to be used for automatic PLAN_TABLE creation. It applies only if the fourth parameter has a value of Y or Yes and there is a database name specified in the fifth parameter.

The DB2EXPLAIN keyword can be repeated for each DB2 subsystem that the customer wants to enable for EXPLAIN requests. To make it easier to enable EXPLAIN requests on all DB2 subsystems, an asterisk can be specified for the subsystem name. In its simplest form, a single entry can be specified as follows:

DB2EXPLAIN=(*)

This specifies that EXPLAIN requests can be run on all DB2 subsystems, with default values for SQL STRING DELIMITER and USE FOR DYNAMICRULES, and without automatic PLAN_TABLE creation.

If automatic PLAN_TABLE creation is required on all DB2 subsystems (for DB2, Version 7 or earlier, or Version 8 in COMPAT mode), and default values aply for SQL STRING DELIMITER and USE FOR DYNAMICRULES, then the following syntax can be used:

DB2EXPLAIN=(*,,,Y)

If the same database name is to be used for automatic PLAN_TABLE creation on all DB2 subsystems, the following syntax can be used:

DB2EXPLAIN=(*,,,Y,dbname)

If the same table space name is to be used for automatic PLAN_TABLE creation on all DB2 subsystems, the following syntax can be used:

DB2EXPLAIN=(*,,,Y,dbname,tsname)

A maximum of 22 DB2 subsystems can be identified by 22 separate DB2EXPLAIN entries. If more than one entry is specified, and one of those has an asterisk for the DB2 subsystem name, then the started task will reject the configuration and fail to start.

DB2I
Optional.

This is used to specify that the DB2 Intercept should be turned on. Specify DB2I=YES to turn on this feature. Turning this feature on makes the DB2+ data extractor available, which allows additional DB2 data to be collected.

Default: NO.

Note:
Running Application Performance Analyzer measurements with the DB2+ data extractor turned on causes each DB2 call to be intercepted to collect additional data. This might have a small impact on the performance of the target address space. Care should be taken when using this feature with other products that also intercept DB2 calls because unpredictable results might occur. You might want to restrict its use in production environments. You can use Access Rules to restrict the DB2+ feature to specific users. See UseDB2Plus in Access rule descriptions.
DB2IMaxTraceSize
Optional.

Specifies the number of DB2+ SQL call interceptions for which full details will be written to the sample file. Collecting full details on every interception allows the F02 Timeline report to report exact times for all SQL calls. The F02 report will be truncated at the number of calls specified in DB2IMaxTraceSize. The DB2+ data extractor continues to collect the data it requires for the other reports for the duration of the measurement. Use this parameter if you need more or less than the default 100000 entries on the F02 report. It has a range of 0 to 9999000, and must be specified in increments of 1000. Larger values will increase report load time and can cause dataspace overflows. This parameter does not affect DDF tracing.

This parameter can also specify whether the maximum trace size field is to be displayed on Panel 2 - Options, where users can change the value when creating or modifying observation requests.

Examples:

Default: (100000,N)

DB2PlanName
Optional.

The plan name must be specified if you want Application Performance Analyzer to gather the necessary data to report plan and package BIND timestamps. You must also bind the Application Performance Analyzer DB2 plan to use this feature.

DB2Qualifier
Optional.

This entry specifies the qualifier to be used for the PLAN_TABLE that is dedicated to the Application Performance Analyzer started task. If omitted, it defaults to the authorization ID of the started task. It should be used only if more than one Application Performance Analyzer started task is running in the same MVS image.

The primary purpose for DB2Qualifier is to prevent using the same PLAN_TABLE for more than one started task. If two started tasks are each processing an EXPLAIN request at the same time using the same PLAN_TABLE, then one will have to wait while the other executes. Giving each started task its own qualifier prevents serializing the EXPLAIN requests among the started tasks.

DefLocSys
Optional.

Specifies to Application Performance Analyzer whether or not the local system id is to be used as the default for NEW observation requests. The acceptable values for this parameter are YES and NO. A value of YES directs Application Performance Analyzer to populate the System Name field on NEW observation requests with the local system id. A value of NO populates the field with *. When the System Id field is set to * and a non-active job is specified, Application Performance Analyzer looks at all systems in the Sysplex Group and starts measuring the first one it encounters. When the System Id field is set to * and an active job is specified, the active job list displays jobs from all systems in the Sysplex Group. In any case, the user can type over the field. This parameter specifies how it should initially be populated.

Default: NO

DescRequired
Optional.

Specifies whether the user is required to complete the description field on NEW, TNEW and MOD requests. YES and NO are the only valid values. YES indicates a description of at least 8 characters is required.

Default: NO

DEXMR
Optional.

Specifies whether the Data Extractor Module Resolution User Exit (CAZDEXMR) is called to resolve PSW addresses that were unresolved by the sampling engine. Specify YES to activate. A PSW address might be unresolved because the address was sampled in a foreign address space, or the address belongs to a module that is not managed by Content Supervision Services (i.e. LOAD, LINK, XTCL, ATTACH).

Default: NO

DSPMaxSize
Optional.

Specifies the maximum amount of storage in 4K blocks that the sampling data space is permitted to use. This storage limit applies to each sampling session that is running. The DSPMaxSize parameter will not directly limit the data space size used for reporting, but will affect reporting by limiting the amount of data that can be collected for reporting.

The minimum amount that can be specified is 30,000, the maximum is 514,288.

Default: 50,000 blocks (approximately 200MB)

DuplicateDSN
Optional.

Specifies the action Application Performance Analyzer is to take when the data set name of the sample file it is attempting to create already exists. This parameter is applicable only when the SampleDSN setting has been specified in CONFIG SAMPLE. The acceptable values are REP, GEN and DIS.

Default: GEN

ExtractorsOnByDefault
Optional.

Specifies defaults for specific data extractors that are on for a NEW command line measurement request. (NEW entered as a line command on an existing request will ignore this parameter, and clone the target request's extractors as usual). The data extractor names are listed separated by commas. Valid values are: CICS, CICS+, DB2, DB2+, DB2V, CDB2, IMS, IMS+, JAVA, and MQ. For example, to have CICS, DB2 and DB2+ turned on by default, code the following:ExtractorsOnByDefault=(CICS,DB2,DB2+)

The parameter AllExtractorsOn, if set to YES, will override this parameter and turn on all extractors.

Default: NO

FillDefaults
Optional.

Specifies whether the installation default values for Number of Samples (NSamples) and Duration (SampleDur) are automatically completed on all NEW and TNEW requests. YES and NO are the only valid values.

Default: NO.

HFSReqLim
Optional.

Specifies the limit for the number of HFS files to be observed in each observation request. HFSReqLim specifies the maximum number of HFS files for the entire observation request. The value must be a number from 1 to 64000. Use this parm if you wish to either limit the number of HFS files to be observed, or increase the number above the default maximum of 500. Measuring more HFS file affects the performance of the measurement.

Default: 500

IMSI
Optional.

This is used to specify that the IMS™ Intercept should be turned on. Specify IMSI=YES to turn on this feature. Turning this feature on makes the IMS+ data extractor available, which allows additional IMS data to be collected.

Default: NO.

Note:
Running Application Performance Analyzer measurements with the IMS+ data extractor turned on causes each IMS call to be intercepted to collect additional data. This might have a small impact on the performance of the target address space. Care should be taken when using this feature with other products that also intercept IMS calls because unpredictable results might occur. You might want to restrict its use in production environments. You can use Access Rules to restrict the IMS+ feature to specific users. See UseIMSPlus in Access rule descriptions.
IMSIMaxTraceSize
Optional.

Specifies the number of IMS+ DLI call interceptions for which full details will be written to the sample file. Collecting full details on every interception allows the I02 and I03 Timeline reports to report exact times for all DLI calls and IMS transactions. The I02 and I03 reports will be truncated at the number of calls specified in IMSIMaxTraceSize. The IMS+ data extractor continues to collect the data it requires for the other reports for the duration of the measurement. Use this parameter if you need more or less than the default 100000 entries on the I02 and I03 reports. It has a range of 0 to 9999000, and must be specified in increments of 1000. Larger values will increase report load time and can cause dataspace overflows. This parameter can also specify whether the maximum trace size field is to be displayed on Panel 2 - Options, where users can change the value when creating or modifying observation requests.

Examples:

Default: (100000,N)

J9VMEnable
Optional.

Specifies whether programs running under Java Version 5 and Java Version 6 can be sampled. Specify YES or 64 if you require this ability. A value of YES launches the 31-bit version of Java in the Application Performance Analyzer started task. A value of 64 launches the 64-bit version of Java in the Application Performance Analyzer started task. If YES or 64 is specified, then both JavaHome and USSHome must also be specified. Refer to Customizing Application Performance Analyzer for use with J9VM for further details.

Note:
Regardless of the whether YES or 64 is specified, Application Performance Analyzer can sample both 31-bit and 64-bit Java V5 and V6 applications. If you have multiple versions of Java installed, only one version can be specified for the Application Performance Analyzer started task at start up, but any version of Java supported by Application Performance Analyzer can be run in the target application.

Default: NO.

JavaHome
Optional.

Specifies the path name to the home directory of the most recent Java installation that supports attaching to a target J9VM. If specified, this should look similar to the following:

JavaHome=/usr/lpp/java/J5.0 

If your system is different, specify the directory path to Java Version 5 or Java Version 6 on your system. This applies only if J9VMEnable=YES or 64. Refer to Customizing Application Performance Analyzer for use with J9VM for further details.

Note:
Ensure you have the correct Java PTF installed on your system to enable Java monitoring for Application Performance Analyzer. Refer to Customizing Application Performance Analyzer for use with J9VM for further details.

Default: none.

MaxMIPPercent
Optional.

Specifies the maximum percentage of MIPS (or machine instructions that a computer can execute in one second) that Application Performance Analyzer will be allowed to consume. The MaxMIPPercent parameter is used to ensure that the total number of MIPS used concurrently by Application Performance Analyzer observation requests does not exceed the specified percentage. When this percentage is reached, new samples created will run at a reduced rate for the entire sample. The reduced rate is shown in S01. If it is already over the limit, new requests will be denied due to a resource shortage. It is recommended that this value be set somewhere between 10 and 50 percent. Typically, a lower value is used on a production machine and a higher value on a development machine.

Default: 25 percent.

MaxSampleRate
Optional.

Specifies the maximum rate at which a measurement session is permitted to sample. When the user submits a measurement request, the user's requested sampling rate is compared to this value for validity. The value is specified in samples per second.

Default: 1,000 samples per second.

MaxSampleSize
Optional.

Specifies the maximum number of samples that a measurement session can record. When the user submits a measurement request, the requested number of samples is compared to this value for validity. The value is specified in samples.

The maximum value you can specify is 175,000.

Default: 30,000 samples.

MLPALibs
Optional.

Application Performance Analyzer can find PLPA library names, but not MPLA library names. This parameter is used to specify up to 10 MPLA loadlib DSNs to be searched for MLPA modules. Add a new MLPALibs parameter for each data set name in the format: MLPALibs=(your.dataset.name)

MultiJobLimit
Optional.

Specifies the limit for the number of jobs allowed in a multiple job request.

Default: 20

Natural
Optional.

Specify YES to make the Natural data extractor available.

Default: NO

NSamples
Optional.

Specifies the default number of samples that a measurement session collects during the measurement session. This value is used if the user omits a value when he creates a measurement request. Normally, the user would enter this value.

Default: 1,000 samples.

SampleDSN
Optional.

Specifies up to 50 characters to be used to name sample data sets. The value in this setting overrides the Application Performance Analyzer generated sample data set name. The name may consist of system symbols and Application Performance Analyzer defined user symbols, which are substituted with values at sample data set allocation time. The total length is restricted to 44 characters after substitution.

The Application Performance Analyzer defined user symbols are:

Note:
System symbol &JOBNAME will substitute the name of the Application Performance Analyzer started task, not the job being sampled.

Example: SampleDSN=SYS2.CAZ0.&SYSNAME..&JOB..D&LYYMMDD..&SAMP

Default: none.

SampleDur
Optional.

Specifies the default length of time that a measurement session collects sampling data. This value is used if the user omits a value when he creates measurement request. The value is specified in seconds. Normally, the user would enter this value.

Default: 60 seconds.

SampleHLQ
Optional.

Specifies the high-level qualifier to be used when creating sample data sets.

SampleHLQ can include system symbols for any qualifier. At sample data set allocation time, the current value of the system symbol will be substituted.

Example when the value of &SYSNAME is SYS1:

SampleHLQ=&SYSNAME..CAZ0     resolves to SYS1.CAZ0 
SampleHLQ=CAZ0.&SYSNAME      resolves to CAZ0.SYS1  

SampleHLQ can also include the keyword &USERID as the first qualifier. At sample data set allocation time the TSO userid of the user requesting the sample will be substituted for this variable.

SampleHLQ must comply with the JCL rule for specifying a data set name. The length of SampleHLQ is still restricted to 17 characters, including the substitutions.

If this parameter is not specified, the value that you specify for HLQ on the CONFIG BASIC statement will be used.

Default: none.

USSHome
Optional.

Specifies the USS directory in which the Application Performance Analyzer USS components have been installed. If specified, this should look similar to the following:

USSHome=#PathPrefix/usr/lpp/apa/v11r1

where #PathPrefix is the appropriate high level directory name. This #PathPrefix string must match the specification for #PathPrefix in the CAZWISMK installation job found in hlq.SCAZSAMP. This applies only if J9VMEnable=YES or 64. Refer to Customizing Application Performance Analyzer for use with J9VM for further details.

Default: none.

USSMaxCM
Optional.

This setting limits the number of USS address spaces that can be measured concurrently. It is intended to control the number of observations that would be initiated due to measured parent processes spawning or forking new processes. It has a range of 1 to 255. It is ignored if USSSpwnEnable=NO.

This is a system-wide limit for all requests initiated by all users.

This setting causes a control block to be allocated in ECSA. The size can be computed as follows: 80 + (MaxUSSCM x 48).

The control block is created once only and remains for the duration of the IPL. Its size can only be altered by re-IPLing the system.

If you change this value and recycle the STC, message CAZ0080W will be issued if the value has been increased. The increased value is ignored.

Default: 48

USSMaxObsPReq
Optional.

This parameter limits the number of child address spaces originating from the target address space that can be measured by a request. In this sense child address spaces are those created using the spawn, fork or exec functions.

A value of zero prevents any child address space from being measured. The maximum value is 255. This value cannot exceed the value specified in USSMaxCM.

The parameter is ignored if USSSpwnEnable=NO.

Default: 10

USSSpwnEnable
Optional.

If set to YES, this setting will enable Application Performance Analyzer to measure USS address spaces that are spun off or substeps that are executed by the original measured (parent) address space.

Default: NO

Note:
USS applications spawn many address spaces. You might need to set USSSpawnEnable to NO if address spaces that are created are not being sampled.
WLMI
Optional.

This is used to specify whether the Application Performance Analyzer WLM Intercept is to be turned on or off. YES and NO are the only valid values. Specify WLMI=YES to turn on this feature. WLMI=YES allows you to measure specific DB2 stored procedures and user-defined functions. It also makes the collateral DB2 (CDB2) data extractor available. When the CDB2 data extractor is selected, Application Performance Analyzer can measure DB2 activity in stored procedures and user-defined functions that are invoked by the application being measured.

When WLMI=YES is specified, Application Performance Analyzer ensures that no other third party product is already intercepting WLM calls. If one is found, a diagnostic error message is displayed, identifying the job name of the third party product that is already intercepting WLM and ignore the WLMI option. To avoid this, Application Performance Analyzer should always be started before any other product that intercepts WLM calls.

Running Application Performance Analyzer measurements with the WLMI data extractor turned on causes each WLM call to be intercepted to collect additional data. This can have a small impact on the performance of the target address space. You might wish to restrict its use in production environments. You can use Access Rules to restrict the measurement of specific stored procedures and user-defined functions, and the use of the CDB2 data extractor to specific users. For additional information, refer to Suggested Application Performance Analyzer rules.

Default: NO

CONFIG LOGFILE statement

Logging records activity within the Application Performance Analyzer started task. Log data will always be written to SYSOUT, but you have the option of also writing this data to log data sets. LOGFILE settings allow you manage these data sets. The following settings might be specified:

Note:
The alphabetical order specified below is not the order of the parameters in the configuration file.
LogGDG
Optional.

If you have specified LogToFile=YES, then LogGDG allows you to specify whether the log data sets will be allocated as GDGs or normal sequential data sets. If you specify LogGDG=YES, you must first create the Generation Data Group Index using an index name of yourhlq.LOG (or LogHLQ.LOG if you choose to use LogHLQ). YES and NO are the only valid values.

Default: NO.

Note:
A GDG log file cannot be shared by multiple instances of Application Performance Analyzer. A second started task with the same GDG log file will fail on start up with messages IKJ56225I and CAZ1030S. A unique GDG log file must be defined for each instance of Application Performance Analyzer.
LogHLQ
Optional.

Specifies the high-level qualifier to be used when creating log data sets. LogHLQ must comply with the JCL rule for specifying a data set name. It can be no more than 17 characters long. A data set will only be created if LogToFile=YES is specified.

If this parameter is not specified, the value that you specify for HLQ on the CONFIG BASIC statement will be used.

Default: none.

LogPRIM
Optional.

If you have specified LogToFile=YES, then LogPRIM allows you to specify the size of each log data set that is allocated. No secondary specification is permitted; therefore, when the log data set has filled, it will be closed and a new log data set will be allocated.

Default: 15.

LogToFile
Optional.

Specifies whether you want log data to be written to permanent data sets. YES and NO are the only valid values.

Default: NO.

Note:
It is recommended that LogToFile be set to YES. The log can be used to help diagnose problems.
LogUNIT
Optional.

If you have specified LogToFile=YES, then LogUNIT allows you to specify whether you want the log data set to be allocated in Tracks or Cylinders. T and C are the only valid values.

Default: T.

CONFIG OPER statement

These settings pertain to operational parameters and should not be changed by the user. These should only be changed on the advice of IBM. Application Performance Analyzer ships with only one parameter specified, which is DiagLevel=1.




Rate this page

[ Top of Page | Previous Page | Next Page | Contents | Index ]