Start Trace (STRTRC)

The Start Trace (STRTRC) command starts traces of original program model (OPM) programs and Integrated Language Environment (ILE) procedures and Java programs (both compiled and JIT). Tracing can be done for multiple jobs using this command. Any number of trace sessions can be started, but active trace session identifiers must be unique across the system. This command can trace call-return flow, data returned by trace points defined in the operating system, component trace information or all three.

The trace session continues until ended with the End Trace (ENDTRC) command or automatically by the watch for trace event functionality. A trace session can be ended from the same job or a different job.

Restrictions:

To use this command, you must have service (*SERVICE) special authority, or be authorized to the Service Trace function of Operating System through iSeries Navigator's Application Administration support. The Change Function Usage (CHGFCNUSG) command, with a function ID of QIBM_SERVICE_TRACE, can also be used to change the list of users that are allowed to perform trace operations.
If you specify a generic user name for the Job name (JOB) parameter, you must have all object (*ALLOBJ) special authority, or be authorized to the Trace Any User function of Operating System through iSeries Navigator's Application Administration support. The Change Function Usage (CHGFCNUSG) command, with a function ID of QIBM_ALLOBJ_TRACE_ANY_USER, can also be used to change the list of users that are allowed to perform trace operations.
A trace cannot be defined to trace all job names and all users.
At least one generic job name must be specified if *YES has been specified for the Restart after next IPL (RESTRIPL) parameter.
Watch-related parameters cannot be specified if *YES is specified for the RESTRIPL parameter.
Only one STRTRC command across the system can be specified to restart after the next IPL.
When the Watched job (WCHJOB) parameter is specified, the issuer of the command must be running under a user profile which is the same as the job user identity of the job being watched, or the issuer of the command must be running under a user profile which has job control (*JOBCTL) special authority. Job control (*JOBCTL) special authority is also required if a generic user name is specified for the WCHJOB parameter.
If you specify a generic user name in the WCHJOB parameter, you must have all object (*ALLOBJ) special authority, or be authorized to the Watch Any Job function of Operating System through iSeries Navigator's Application Administration support. The Change Function Usage (CHGFCNUSG) command, with a function ID of QIBM_WATCH_ANY_JOB, can also be used to change the list of users that are allowed to start and end watch operations.
You must have operational (*OBJOPR) and execute (*EXECUTE) authorities to the user exit program if specified in Trace program (TRCPGM) parameter, and execute (*EXECUTE) authority to the library where the program is located.
You must have use (*USE) authority to the message queues specified in Watched message queue (WCHMSGQ) parameter, and use (*USE) authority to the library where the message queue is located.

Parameters

Keyword	Description	Choices	Notes
SSNID	Session ID	Name, *GEN	Required, Positional 1
JOB	Jobs	Single values: * Other values (up to 8 repetitions): Element list	Optional
	Element 1: Job name	Qualified job name
	Qualifier 1: Job name	Generic name, name, *ALL
	Qualifier 2: User	Generic name, name, *ALL
	Qualifier 3: Number	000001-999999, *ALL
	Element 2: Thread ID to include	Values (up to 20 repetitions): Character value, ALL, INITIAL, *SELECT
JOBTYPE	Job types	ALL, ACTIVE, *NEW	Optional
MAXSTG	Maximum storage to use	1024-4000000, 10000	Optional
TRCFULL	Trace full	STOPTRC, WRAP	Optional
JOBTRCTYPE	Trace type	Values (up to 2 repetitions): ALL, FLOW, DATA, TRCTYPE	Optional
TRCTYPE	Trace type	Single values: NONE Other values (up to 50 repetitions): Element list*	Optional
	Element 1: Component	Character value, ASPMGT, CLUSTER, CMTCTL, DBHSVR, DBSQL, DEVCFG, DIRSRV, EDRSQL, ENVVAR, FILESVR, FLIGHTRCD, HPT, HTTP, IFS, IPC, IPPDRV, LOCKSPACE, NETSVR, NPS, OPASSIST, OPENSSL, POSIXMISC, POSIXPROC, PTHREAD, QNTC, QSHELL, RAS, REGFAC, SECURITY, SERVICE, SOCKETS, TCPIPCFG, THREADMGT, WRKMGT
	Element 2: Trace level	ERROR, INFO, *VERBOSE
TRCFTR	Trace filter	Name, *NONE	Optional
RESTRIPL	Restart after next IPL	NO, YES	Optional
WCHMSG	Watch for message	Single values: NONE Other values (up to 5 repetitions): Element list*	Optional
	Element 1: Message identifier	Name
	Element 2: Comparison data	Character value, *NONE
	Element 3: Compare against	MSGDTA, FROMPGM, *TOPGM
WCHMSGQ	Watched message queue	Values (up to 3 repetitions): Element list	Optional
	Element 1: Message queue	Single values: SYSOPR, JOBLOG, HSTLOG Other values: Qualified object name*
	Qualifier 1: Message queue	Name
	Qualifier 2: Library	Name, *LIBL
WCHJOB	Watched job	Single values: * Other values (up to 5 repetitions): Element list	Optional
	Element 1: Job name	Qualified job name
	Qualifier 1: Job name	Generic name, name
	Qualifier 2: User	Generic name, name
	Qualifier 3: Number	000001-999999, *ALL
WCHLICLOG	Watch for LIC log entry	Single values: NONE Other values (up to 5 repetitions): Element list*	Optional
	Element 1: Major code	Character value, *ALL
	Element 2: Minor code	Character value, *ALL
	Element 3: Comparison data	Character value, *NONE
WCHTIMO	Length of time to watch	1-43200, *NOMAX	Optional
TRCPGM	Trace program	Single values: NONE Other values: Qualified object name*	Optional
	Qualifier 1: Trace program	Name
	Qualifier 2: Library	Name, *LIBL
TRCPGMITV	Time interval	1-9999, *NONE	Optional

Top

Job name (JOB)

Specifies which jobs are to be traced.

Single values

*: Only the job that issues the STRTRC (Start Trace) command is to be traced.

Element 1: Job name

Qualifier 1: Job name

generic-name: Specify the generic name of the jobs to be traced. A generic name is a character string of one or more characters followed by an asterisk (*); for example, ABC*. The asterisk substitutes for any valid characters. A generic job name specifies all jobs with job names that begin with the generic prefix.
*ALL: All jobs names with the specified job user name are traced. *ALL for the job name is considered to be a generic job specification because it will trace all jobs that meet the job user name qualifiers that you specified.
name: Specify the name of the job to be traced. Up to eight job names may be specified, but only one generic job name is allowed.

Qualifier 2: User

generic-name: Specify the generic user name of the jobs from which trace records are to be collected. A generic name is a character string of one or more characters followed by an asterisk (*); for example, ABC*. The asterisk substitutes for any valid characters. A generic user name specifies all jobs with user names that begin with the generic prefix.
*ALL: All job user name with the specified job name are traced. *ALL for the job user name is considered to be a generic job specification because it will trace all jobs that meet the job name qualifiers that you specified.
name: Specify the user name of the job to be traced.

Qualifier 3: Number

*ALL: All jobs with the specified job name and user name are traced. *ALL for the job number is considered to be a generic job specification because it will trace all jobs that meet the job name and job user name qualifiers that you specified.
000001-999999: Specify the job number to further qualify the job name and user name. You cannot specify a job number if a generic job name qualifier or generic user name qualifier is specified.

Element 2: Thread ID to include

*ALL: All threads within the specified job name and user name are traced.
*INITIAL: Only the initial thread within the specified job name and user name is traced.
*SELECT: A list of thread identifiers is shown from which the user can select up to twenty to be traced. This value is only valid if the command is run in an interactive job.
thread-identifier: Specify the identifiers of up to twenty threads whose information is to be included.

Trace type (TRCTYPE)

Specifies the component to trace and the level to use. You can specify up to 50 trace components to be traced.

Single values

*NONE: No component trace is specified.

Element 1: Component

Each trace type is identified by a special value. Specify trace component special values from the following table:

TRACE DESCRIPTION                  SPECIAL VALUE
ASP Management                     *ASPMGT
Cluster Resource Services          *CLUSTER
Commitment Control                 *CMTCTL
Database Host Server               *DBHSVR
Database SQL                       *DBSQL
Device Configuration               *DEVCFG
Directory Services                 *DIRSRV
Extended Dynamic Remote SQL (EDRS) *EDRSQL
Environment Variable APIs          *ENVVAR
File Server                        *FILESVR
Flight Recorder                    *FLIGHTRCD
Host Print Transform               *HPT
HTTP Server powered by Apache      *HTTP
Integrated File System             *IFS
Interprocess Communication APIs    *IPC
Internet Print Protocol Driver     *IPPDRV
Lock Space Management              *LOCKSPACE
iSeries NetServer                  *NETSVR
Network Print Server               *NPS
Operational Assistant              *OPASSIST
OpenSSL APIs                       *OPENSSL
POSIX Miscellaneous APIs           *POSIXMISC
POSIX Process-Related APIs         *POSIXPROC
POSIX Thread APIs                  *PTHREAD
QNTC File System                   *QNTC
Qshell                             *QSHELL
Remote Access Services             *RAS
Registration Facility              *REGFAC
Software Security                  *SECURITY
Service Facility                   *SERVICE
Sockets APIs                       *SOCKETS
TCP/IP Configuration               *TCPIPCFG
Thread Management                  *THREADMGT
Work Management                    *WRKMGT

Element 2: Trace level

Specifies the trace level to be used for the specified component.

*ERROR: The trace level ERROR will be used. Components typically trace error return codes, exception conditions, and invalid input. The amount of data is usually small.
*INFO: The trace level INFO will be used, which also includes ERROR trace level data. Components typically trace entry and exit from interfaces, parameters and return codes, and major changes of flow or semantics caused by input or other decisions. The amount of data is moderate.
*VERBOSE: The trace level VERBOSE will be used, which also includes INFO and ERROR trace level data. Components typically trace detailed data that could assist in debugging control flow, data corruption, data structures, environment, call stacks, and resource allocations. The amount of data can be large.

Restart after next IPL (RESTRIPL)

Specifies if the trace will be restarted after the next initial program load (IPL) of the operating system.

Note: Only one STRTRC command can be specified to restart after the next IPL. If another STRTRC command was previously specified with RESTRIPL(*YES), an error message is sent and this STRTRC command fails. This parameter can be specified only if at least one generic job name is specified in the Job name (JOB) parameter. You must run the End Trace (ENDTRC) command for the trace session which specified RESTRIPL(*YES) to clear the storage used to hold the STRTRC command string to be restarted after an IPL. This parameter cannot be specified if any watch-related parameters have been specified.

Note: If the IPL is associated with an install or upgrade of the operating system, the trace will not be restarted.

*NO: The trace will not be restarted after an IPL.
*YES: The trace will be restarted after the next IPL. The trace being started by this STRTRC automatically ends if an IPL of the operating system is performed. The trace started after the next IPL will be a new trace session with new trace data. You can use the ENDTRC command to end the trace.

Watch for message (WCHMSG)

Specifies up to five message identifiers which are to be watched for. If a value other than *NONE is specified, you must specify where to watch for the message on the WCHMSGQ parameter. When the watched for message is added to the specified message queue or log, the trace exit program is called; if no trace exit program is defined, the trace stops.

Single values

*NONE: No messages will be watched for.

Element 1: Message identifier

name: Specify the 7-character message identifier to be watched for.

Element 2: Comparison data

Specify comparison data to be used if a message matching the specified message ID is added to the specified message queue or log. If the message data, the "From program" or the "To program" includes the specified text, the watched for condition is true. If the message data, the "From program" or the "To program" does not contain the specified text, the trace function continues.

*NONE: No comparison data is specified. If a message matching the specified message ID is added to the specified message queue or log, the watched for condition is true.
character-value: Specify the text string used to compare against the message data, the "From program" or the "To program" of the watched for message. This text is case sensitive and can be quoted in order to specify imbedded or trailing blanks.

Element 3: Compare against

Specify which part of the message the comparison data specified for element 2 is to be compared against.

*MSGDATA: The comparison data will be compared against the message replacement data.
*FROMPGM: The comparison data will be compared against the name of the program sending the message, or the name of the ILE program that contains the procedure sending the message.
*TOPGM: The comparison data will be compared against the name of the program the message was sent to, or the name of the ILE program that contains the procedure the message was sent to.

Watched message queue (WCHMSGQ)

Specifies where to watch for the message identifiers specified on the WCHMSG parameter. You can specify to watch the message being added to the system operator message queue, the history log, other message queues, and job logs. Up to three message queues or special values can be specified.

Element 1: Message queue

Single values

*SYSOPR: Watch messages added to the system operator message queue (QSYSOPR message queue in library QSYS).
*JOBLOG: Watch messages added to the job logs of the jobs specified for the Watched job (WCHJOB) parameter.
*HSTLOG: Watch messages added to the history log (QHST message queue in library QSYS).

Qualifier 1: Message queue

name: Specify the name of the message queue to watch.

Qualifier 2: Library

*LIBL: All libraries in the library list for the current thread are searched until the first match is found.

name: Specify the name of the library where the message queue is located.

Watched job (WCHJOB)

Specifies the job whose job log is watched for the messages specified on the WCHMSG parameter. The specified job will only be watched if *JOBLOG is specified on the WCHMSGQ parameter. Up to five job names may be specified.

Single values

*: Only the job log of the job that issued this trace command is watched.

Element 1: Job name

Qualifier 1: Job name

generic-name: Specify the generic name of the job to be watched. A generic name is a character string of one or more characters followed by an asterisk (*); for example, ABC*. The asterisk substitutes for any valid characters. A generic job name specifies all jobs with job names that begin with the generic prefix.
name: Specify the name of the job to be watched.

Qualifier 2: User

generic-name: Specify the generic name of the user name of the job to be watched. A generic name is a character string of one or more characters followed by an asterisk (*); for example, ABC*. The asterisk substitutes for any valid characters. A generic user name specifies all jobs with the specified job name and with user names that begin with the generic prefix.
name: Specify the user name of the job to be watched.

Qualifier 3: Number

*ALL: All jobs with the specified job name and user name are watched.
000001-999999: Specify the job number to further qualify the job name and user name. You cannot specify a job number if a generic job name or a generic user name qualifier is specified.

Watch for LIC log entry (WCHLICLOG)

Specifies up to five licensed internal code (LIC) log entry identifiers which are to be watched for. Each LIC log entry contains a major and a minor code. The watched for condition will be met if a LIC log entry is added that matches the specified major and minor codes and any comparison data specified. When the watched for log entry is added to the LIC log, the trace exit program is called, even when the comparison data specified does not match; if no trace exit program is defined, the trace stops.

Single values

*NONE: No LIC log entries will be watched for.

Element 1: Major code

*ALL: Any LIC log entry major code will be considered to be a match. If *ALL is specified for the major code, you cannot specify *ALL for the LIC log entry minor code.
character-value: Specify the LIC log major code to be watched for. You can specify either a hexadecimal digit or a question mark for each character in the four-digit code. A question mark is a wildcard character that will match any digit in that position. Up to three wildcard characters can be specified.

Element 2: Minor code

*ALL: Any LIC log entry minor code will be considered to be a match. If *ALL is specified for the minor code, you cannot specify *ALL for the LIC log entry major code.
character-value: Specify the LIC log minor code to be watched for. You can specify either a hexadecimal digit or a question mark for each character in the four-digit code. A question mark is a wildcard character that will match any digit in that position. Up to three wildcard characters can be specified.

Element 3: Comparison data

Specify comparison data to be used if a log entry matching the specified major and minor codes is added to the licensed internal code (LIC) log. If this text is found in the LIC log entry data fields of the watched for log entry, the watched for condition is true. If this text is not found in the LIC log entry data fields of the watched for log entry and no exit program is specified on the TRCPGM parameter, the trace function continues. If the log entry matches the specified major and minor codes and an exit program is specified on the TRCPGM parameter, but the entry data does not contain the specified text, the exit program is called to determine if the trace should continue or stop.

*NONE: No comparison data is specified. If a LIC log entry matching the specified major and minor codes is added to the LIC log, the watched for condition is true.
character-value: Specify the text string used to compare against the entry data of the watched for log entry. If this text is found in the LIC log entry data fields compared of a watched for log entry, the watch condition is considered to be true. This text is case sensitive. The LIC log fields which can be compared are TDE number, task name, server type, job name, user ID, job number, thread ID, exception ID, LIC module compile timestamp, LIC module offset, LIC module RU name, LIC module name, LIC module entry point name. The comparison data cannot be used to match across two fields, and can match an entire field or a substring of any field.
When watching for an exception ID, all four hexadecimal digits of the exception ID must be specified. Also, the prefix MCH may be specified if you want to compare only against the exception ID field and avoid possible substring matches with the other fields.

Trace program (TRCPGM)

Specifies the program to be called for user-defined trace commands and procedures.

The trace program will be called:

Before the application trace starts.
After a match of a message identifier specified for the WCHMSG parameter, or a match of a Licensed Internal Code (LIC) log entry specified for the WCHLICLOG parameter occurs.
When the time interval specified on the TRCPGMITV parameter is reached.
When the length of time to watch specified on WCHTIMO parameter is reached.

There are three input parameters and one output parameter associated with the trace program. The four parameters are required:

1   Trace option setting   Input    Char(10)
2   Reserved               Input    Char(10)
3   Error detected         Output   Char(10)
4   Comparison data        Input    Char(*)

Allowed values for the "Trace option setting" parameter are:

*ON: The watch for trace facility is starting when the collection of trace information is started.
*MSGID: A match on a message id specified on WCHMSG parameter occurred.
*LICLOG: A match on a LIC log specified on the WCHLICLOG parameter occurred.
*CMPDATA: The major and minor code of a LIC log matched, but the comparison data did not.
*INTVAL: The time interval specified on TRCPGMITV parameter is elapsed.
*WCHTIMO: The length of time to watch specified on WCHTIMO parameter is elapsed.

The "Reserved" parameter must be set to blanks.

Allowed values for the "Error detected" parameter are:

*CONTINUE: The trace and the watch for trace event facility will continue running.
*STOP: The trace and the watch for trace event facility will be ended.
*ERROR: Error detected by customer trace program.

Allowed values for the "Comparison data" parameter when *MSGID is specified for the "Trace option setting" parameter will be the following structure:

OFFSET     TYPE         FIELD
Dec Hex
 0   0     BINARY(4)    Length of trace information
 4   4     CHAR(7)      Message ID
11   B     CHAR(9)      Reserved
20  14     BINARY(4)    Offset to comparison data
24  18     BINARY(4)    Length of comparison data
 *   *     CHAR(*)      Message comparison data

Allowed values for the "Comparison data" parameter when *LICLOG or *CMPDATA is specified for the "Trace option setting" parameter will be the following structure:

OFFSET     TYPE         FIELD
Dec Hex
 0   0     BINARY(4)    Length of trace information
 4   4     CHAR(4)      LIC Log major code
 8   8     CHAR(4)      LIC Log minor code
12   C     CHAR(8)      LIC Log identifier
20  14     BINARY(4)    Offset to comparison data
24  18     BINARY(4)    Length of comparison data
 *   *     CHAR(*)      LIC log comparison data

Allowed values for the "Comparison data" parameter when *ON, *INTVAL or *WCHTIMO is specified for the "Trace option setting" parameter will be the following structure:

OFFSET     TYPE         FIELD
Dec Hex
 0   0     BINARY(4)    Length of trace information (always 4).

For more information on the trace exit program interface, refer to the System API Reference information in the iSeries Information Center at http://www.iseries.ibm.com/infocenter .

Single values

*NONE: No trace exit program is defined. If a watched for message or licensed internal code (LIC) log entry is added, or if the specified watch time limit is exceeded, the trace function ends.

Qualifier 1: Trace program

name: Specify the name of the trace exit program.

Qualifier 2: Library

*LIBL: All libraries in the job's library list are searched until the first match is found.
name: Specify the name of the library where the user exit program is located.

Examples

Example 1: Start a Trace on Your Job

STRTRC   SSNID(*GEN)

This command starts a trace on the job that called the STRTRC command. A unique trace session identifier will be generated. The session identifier will be returned in the message data of a completion message sent after the trace session starts successfully.

Example 2: Start a Trace of Another Job

STRTRC   JOB(123456/QSYS/QSYSARB)  SSNID(ARBTRACE)

This command starts a new trace on the job with job name QSYSARB, job user name of QSYS, and job number of 123456. Only this one job is traced. The trace session identifier is ARBTRACE.

Example 3: Start a Generic Job Trace

STRTRC   SSNID(*GEN)  JOB(*ALL/QSYS/QTVTELNET*)  MAXSTG(80000)

This command starts a new trace on all jobs on the system that have a job name that begins with QTVTELNET and are running under the QSYS user profile. The maximum storage used for the trace will be 80000 kilobytes. A unique session identifier will be generated.

Example 4: Start a Job Trace with a Filter

ADDTRCFTR   FTR(MYFTR)  PGMTRG(*EQ MYLIB/MYPGM *ALL *PGM *ENTRY)
STRTRC   SSNID(MYTRACE)JOB(123456/MYUSER/MYJOB)  TRCFTR(MYFTR)

These commands add a trace filter for MYPGM in MYLIB. When the trace is started, call/return trace records will not be collected until MYLIB/MYPGM is called. Once MYLIB/MYPGM is called, the trace will collect call/return trace records until the trace is ended.

Example 5: Start a Job Trace with Component Defined

STRTRC   JOB(123456/MYUSER/MYJOB)  SSNID(MYTRACE)
         TRCTYPE(*ENVVAR *VERBOSE)

This command starts a new trace on the job with job name MYJOB, job user name of MYUSER, and job number of 123456. Only this one job is traced. The trace session identifier is MYTRACE. The trace will collect call/return trace information as well as information for the ENVVAR component at VERBOSE level.

Example 6: Start a Job Trace with Threads Selection

STRTRC   JOB((123456/MYUSER/MYJOB (11111111 22222222)))
         SSNID(MYTRACE)

This command starts a new trace on the job with job name MYJOB, job user name of MYUSER, and job number of 123456. Only threads 11111111 and 22222222 are traced. The trace session identifier is MYTRACE.

Example 7: Start a Trace and Watch for a Message to End the Trace

STRTRC   SSNID(*GEN)  WCHMSG((MCH2804))
         WCHMSGQ((*SYSOPR) (*JOBLOG))
         WCHJOB((*ALL/MYUSER/MYJOBNAME))
         TRCPGM(MYLIB/TRCEXTPGM)

This command starts a trace on the job that called the STRTRC command. The trace will be ended when MCH2804 message is found on the System Operator message queue or within the *ALL/MYUSER/MYJOBNAME job log. Also, MYLIB/TRCEXTPGM is specified as a trace exit program.

Example 8: Start a Trace and Watch for a LIC Log Entry to End the Trace

STRTRC   SSNID(*GEN) WCHLICLOG(('99??' 9932 MYJOBNAME))
         WCHTIMO(*NOMAX)

This command starts a trace on the job that called the STRTRC command. The trace will be ended when a Licensed Internal Code (LIC) log entry that has a major code starting with 99 and a minor code of 9932 is generated on the system. Also, the LIC log information should contain the text "MYJOBNAME". *NOMAX on WCHTIMO parameter indicates that the trace will be active until the event occurs or ENDTRC command is issued manually.