Com-plete capture processing is a mechanism for the "capturing" and saving of data of any type for later use. The data is captured to socalled CAPTURE data sets which must be allocated and initialized before Com-plete is initialized. As well as capturing system data, such as the trace records, installation-specific data can also be captured.
This document covers the following topics:
The Capture data sets are VSAM data sets allocated as specified in the section Com-plete Files and Associated User Files. They must then be included in the Com-plete job control with the DDname/DLBL name CAPTURn, where n is a number from 1 to 9. At initialization, capture processing either uses the number of Capture data sets as specified in the CAPTURE sysparm, or determines the number of data sets specified in the job control dynamically.
Note:
To disable Capture, you must specify SUBSYS-IGNORE=CAPTURE in the
Com-plete sysparms.
By default, the capture data sets are non-reusable. This means that when they are filled, they cannot be used again until they have been copied and/or reset by the TUSACAPT utility. To enable the capture subsystem to reuse the data sets when they are filled, specify CAPTURE=(n,REUSE) in the sysparms.
After initialization, Capture opens the first available data set and uses it to capture data. The capture data currently in use is always clearly visible from the COMCA... messages issued to the operator console. When a data set fills, the capture subsystem informs the operator and attempts to find another data set that can be used. If the REUSE option is specified on the CAPTURE sysparm, the next data set is selected, opened and used (in round robin fashion, from CAPTUR1 to CAPTUR9), regardless of whether it already contains data or not.
If the REUSE option is not specified, and the current data set
fills, the next data set is checked for data. If data exists, the following
data set is checked, and so on, until a data set without data is found. In the
event that all data sets contain capture data, a message goes out to the
operator, informing him/her to clear a data set for use, otherwise capture
processing cannot continue. Com-plete then waits until the operator replies
that one or more data sets have been cleared before checking again for an
'empty' data set. For this reason, it is recommended that you build a procedure
to unload and initialize Capture data sets as soon as they are filled.
When a Capture data set is opened, either for the first time or after it has been unloaded and initialized, a header record is written as the first record to the Capture data set. This contains details of times, dates, software levels etc. of when the data set was opened. The exact contents and format of this header record can be found in the CAPLAB macro supplied on the distributed Com-plete source data set.
Following this header, system and/or user capture records can be found. Each record has a prefix to identify the type and format of the following record along with the date and time it was written, followed by the actual data itself. Refer to the CCCAPT copy book supplied on the Com-plete distributed source for details of the layout of this area.
Com-plete provides no utilities to process these records, as the contents of the records is entirely up to the user.