The trace output data sets must be specific to each instance of
GTF and can be defined in the cataloged procedure. Each instance of
GTF to be started must have a separate cataloged procedure, or if
the same cataloged procedure is used, then a different trace data
set must be supplied with the START command
Use the following guidelines for specifying trace output data sets
on the IEFRDER DD or GTFOUTxx DD statements:
- You can define up to 16 output data sets for GTF to use. If you
define more than 16 data sets, GTF will accept the first 16 and ignore
the rest.
- If GTF cannot open all of the data sets, it issues a message that
identifies those that are unopened, and continues processing with
those that are open.
- Do not specify the RLSE option while using the SPACE parameter
because the output data sets are opened and closed more than once
while GTF runs.
Note: If the GTF trace output resides on a SMS managed
volume you should ensure the SMS management class does not allow partial
release.
- Do not request secondary extents for trace data sets. GTF will
only use the first extent.
To obtain the maximum degree of control
over the number of trace entries for which space is allocated, specify
space allocation in blocks, using a block length that matches the
BLKSIZE of your trace data set. Do not specify any secondary space.
Use the CONTIG option to request contiguous blocks. For example, if
your BLKSIZE is 8192, code the SPACE keyword as follows:
SPACE=(8192,(500,0),,CONTIG)
- All data sets must be in the same device class: either DASD or
tape, but not both. If you mix device classes, GTF will ignore the
tapes and use only DASD. However, the data sets can have different
device types; for example, you can mix 3380 and 3390 device types.
- DSNTYPE=LARGE data sets are treated like conventional data sets
except they can occupy more than 64K tracks per volume.
- VSAM linear data sets are treated like conventional data sets.
The use of a data set is rejected if it does not comply to the 32K
control interval size requirement. The VSAM extended addressing option
is accepted.
When WRAP processing is requested, the primary space
request is consulted and only control intervals contained within that
space are used. Unlike non-VSAM data sets used for WRAP processing,
the primary space does not have to be satisfied using a single extent. A
data set must be empty or have been defined with the REUSE attribute.
If neither of the two conditions exists, GTF rejects the use of the
data set.
When WRAP processing is not requested, control
intervals are filled until GTF is stopped or the data set is full.
Note: GTF
and CTRACE accept a single VSAM linear data set as output. VSAM's
support for striping can increase data rate without the complexity
associated with the use of distinct data sets.
- GTF and CTRACE support placement of NOWRAP traces in cylinder-managed
space. WRAP traces placed in VSAM linear data sets can reside in cylinder-managed
space too. WRAP traces in non-VSAM data sets cannot be placed in large
format data sets, extended format data sets, or cylinder-managed space.
- To ensure the most efficient GTF processing, do not specify any
particular block size for the output data set or data sets in either:
- The cataloged procedure for GTF
- The JCL, TSO/E commands, or interactive system productivity facility/program
development facility (ISPF/PDF) panels that you might use to preallocate
the data set or data sets
The system computes an optimal block size when it opens each
data set. EXCEPTION: If you want GTF to use an unlabeled
tape as the output data set, you must specify the logical record length
and block size when you allocate that data set.
- If you define more than one data set, you should ensure that the
number of paths to the data sets equals the number of data sets.
- You can specify the number of channel programs for each output
data set using the NCP parameter on each DD statement. The NCP value
determines the rate at which GTF transfers data to the output data
sets. For example, if you want to transfer data to your data sets
at a rate of 25 buffers per second and you have 5 data sets, you will
need to specify an NCP value of 5. GTF then transfers data to the
5 data sets at a rate of 5 buffers per second per data set for a total
rate of 25 buffers per second.
The maximum value for NCP is 255.
If you do not specify a value for NCP, or if you specify a value less
than four, GTF will use the following default values:
- For tape: four
- For DASD: the number of output blocks per track, multiplied by
four.
- If, when you enter the START command, you override any of the
DD statements for multiple output data sets, you must use symbolic
parameters in those DD statements. See Using the START command to invoke GTF for
more information.