Specifying TRACE TT options
- COLL | C=collection
- specifies a collection name to be used as a filter for tracing.
Collection name identifies the customer-defined name for a group of
associated packages.
The collection name consists of 1 through 18 characters.
- CON=connection type
- specifies a connection type to be used as a filter for tracing.
Connection type identifies the type associated with the environment
creating the work request, which may reside anywhere within the network.
The connection type consists of 1 through 8 characters.
- COR=correlation info
- specifies correlation information to be used as a filter for tracing.
Correlation information identifies the name associated with the user
or program creating the work request, which may reside anywhere within
the network.
The correlation information consists of 1 through 18 characters.
- LU=logical unit name
- specifies the LUNAME to be used as a filter, allowing work related
to this logical unit to be traced.
The luname consists of 1 through 8 characters.
- LVL=level
- specifies the trace level indicator to be assigned to this filter
set. Valid trace level indicators are 1, 2, 3, and 128-255.
The level consists of 1 through 3 numeric characters.
The default trace level value is 2.
Trace level indicators 0, and 4 through 127 are reserved.
- NET=netid
- specifies a NETID name to be used as a filter, allowing work related
to this network ID to be traced.
The netid consists of 1 through 8 characters.
Specifying NET requires the specification of the LU keyword.
- PKG=package
- specifies a package name to be used as a filter for tracing. The
package name identifies a set of associated SQL statements.
The package name consists of 1 through 8 characters.
- PLAN|PL=plan name
- specifies a name to be used as a filter for tracing. The plan
name identifies the access plan name for a set of associated SQL statements.
The plan name consists of 1 through 8 characters.
- PROC|PR=proc name
- specifies the proc name to be used as a filter for tracing. Proc
name identities the DB2-stored SQL Procedure name associated with
the work request.
The proc name consists of 1 through 18 characters.
- PRS=process
- specifies process information to be used as a filter for tracing.
Process information identifies the process name associated with the
work request.
The process information consists of 1 through 32 characters.
- SUB=subsystem
- specifies subsystem information to be used as a filter for tracing.
Subsystem information identifies character data related to the work
request that is passed by the work manager for use in WLM classification.
The subsystem information consists of 1 through 18 characters.
- TC=transaction class
- specifies the transaction class to be used as a filter for tracing.
Transaction class identifies a class name within the subsystem.
The transaction class information consists of 1 through 8 characters.
- TRAN|T=transaction_name
- specifies a transaction name to be used as a filter for tracing.
Although the transaction name has specific meaning to many subsystems (those that exploit WLM classify), the transaction name identifies the name of the work request desired to be traced across the set of components that may handle it. See z/OS MVS Planning: Workload Management for information on transaction name and jobname.
The transaction name consists of 1 through 8 characters.
- USER|U=userid
- specifies a userid to be used as a filter for tracing.
userid is the name of a single userid for whom the tracing is to be done.
The userid consists of 1 through 8 characters.
- WTR=membername|STOP
- indicates that the external writer is to be started for recording
transaction trace entries in a data set, or stopped if it was already
started. This is an optional keyword.
- membername
- identifies the member that contains the source JCL that invokes
the component trace external writer. The system opens the data sets
that the writer uses. The member can be a SYS1.PROCLIB catalogued
procedure or a job.
If WTR= is not specified, the trace data is saved in the TTRACE data space and can be obtained by dumping the data space.
The external writer may be started on the first invocation of the TRACE TT command or on any subsequent TRACE TT command prior to the TRACE TT,OFF command. Only one external writer may be started on any instance of transaction trace.
The external writer proc name must be the same on every system in the sysplex that is participating in transaction trace. Each system may define a proc with the same name or share a common writer proc.
If trace is already running, the writer may be started with a TRACE TT,WTR= command without specifying any filter parameters.
Trace entries are recorded in the external writer data set from the time the external writer is started until the external writer is stopped with a TRACE TT,WTR=STOP or trace is turned off with a TRACE TT,OFF,ALL. When the end of the data set is reached, it starts writing from the top of the data set, overlaying the oldest entries. Specifying a "membername" when transaction trace is not active (that is, no filter has been specified) results in an error message.
- STOP
- indicates that the external writer that was previously started
needs to be stopped. No more trace entries are recorded in the external
writer data set.
Any trace records currently in the buffers are written prior to stopping the external writer. An error message is issued if WTR=STOP is issued when no external writer is active.
- BUFSIZE|BUFSIZ=nnnK or nnM
- specifies the size of the TTrace data space the system is to use,
in kilobytes (K) or megabytes (M).
nnnK is the buffer size in kilobytes, where nnn is a decimal number from 16 to 999.
nnM is the buffer size in megabytes, where nn is a decimal number from 1 to 32.
The buffer size may be changed anytime when trace is active. This includes when the initial TRACE TT command is issued. Specifying a buffer size when transaction trace is not active (that is, no filter has been specified) results in an error message. The default size is 1M.
The scope of this buffer size change affects TTrace processing until the buffer size is changed on a subsequent command or after a system IPL. That is, the buffer size persists if TTrace is turned off and turned on at a later time.
The buffer size is rounded up to a page boundary (that is, multiples of 4K).
This is an optional keyword.
- LATENT=(Y|N)
- specifies whether latent transactions need to be traced. A latent transaction is defined by all of the following:
- The transaction is currently active in the system.
- The transaction is marked for tracing.
- The filter value used to mark the transaction eligible for tracing is no longer active, that is, TRACE TT,OFF=nn was issued to turn that filter off.
Y specifies that latent transactions, if any, can continue to be traced. Y is the default value.
N specifies to not trace latent transactions. Any current active transactions that do not have an active filter will not be traced from this point on.
The scope of the latent parameter setting affects TTrace processing until the parameter is changed on a subsequent command or after a system IPL. That is, the parameter value persists if TTrace is turned off and then turned on at a later time.
- OFF=(nn|ALL)
- turns the specified active transaction trace OFF across the sysplex.
- nn
- specifies the filter set number of the trace to be turned off.
nn is an integer from 1-99 that identifies a particular transaction trace filter set. Use DISPLAY TRACE,TT to display the active TTrace filter sets.
- ALL
- specifies that all the active transaction traces should be turned
off across the sysplex.
No additional trace entries are recorded. Any data in the trace buffers is written out to the data space and external writer data set. The external writer, if it was started, is also stopped.
A message is issued to the operator console to indicate that transaction trace is no longer active on the system and a DUMP command may be issued to access the trace in the data space.