Define logs in your initialization file with ALOG1 and ALOG2 statements. CA WA CA 7 Edition writes many different log records to the log files to track its activity. As a best practice (to help ensure no log records are lost), primary and secondary logs are defined in the initialization file with the ALOG1 and ALOG2 statements. The CA7ONL procedure has a DD for the primary log with the ddname UCC7LOG. This DD statement causes the allocation of the primary log file. You can use a DD statement for the secondary log (you can call it UCC7LOGS) to allocate this log file also. This naming helps ensure that neither of the log files is moved during any DFSMS utility operations.
The DBASE statement in the initialization file JOB= keyword names a 7-character log dump job. The job is automatically submitted when the log fills to archive that data while recording to the alternate log file. A P or an S is appended to the seven character job name to denote the primary (P) or secondary (S) file to be dumped. The default job names are CA07LOGP and CA07LOGS. These jobs execute program SASSHIS5, which writes the log records to a history file.
These log dump jobs are expected to complete typically within CA WA CA 7 Edition control. If there are problems with the log dump jobs, such as abends or being canceled, then the message CA-7.200 BOTH LOG FILES FULL is issued. Because CA7ONL creates many log records, normal processing is suspended when the WTOR is issued. Jobs are not submitted to the request queue. A user cannot log on. Jobs do not move between queues, and so forth. CA7ONL tracks the log file status in the checkpoint data set. CA7ONL determines which log file is currently being written and the successful completion of the last log dump job. This knowledge is why the successful completion of log dump jobs is important.
If you receive the CA-7.200 WTOR, the first step is to try to execute one of the log dump jobs outside of CA7ONL (because it has suspended processing). If the log dump job is successful, reply that file identifier to the WTOR (LOGP for CA07LOGP or LOGS for the CA07LOGS job). CA7ONL then starts recording on the file specified and resets the checkpoint to show the new information.
Business Value:
The capability to recall log records when needed is critical for reporting purposes and in the event product or production troubleshooting is required.
Additional Information:
The SASSHIS5 program can also remove log records from the log history file and place them on the log archives file so the log history file does not become too large. This transfer can occur either while the log records are being extracted from the log files or as a separate operation. Make the DD for the log files a DD DUMMY, and use the PARM operand on the EXEC statement.
To archive records from the history file as a separate operation, have a job (other than the CA07LOGP or CA07LOGS) set up to do so. The other job is required because the log dump jobs cannot be DEMANDed into CA7ONL. They only automatically come into the queues when a log fills or you do the /SWAP command in CA WA CA 7 Edition). The PARM operand on the EXEC statement for the SASSHIS5 program has the following format:
PARM='yyddd,nnnnnn'
yyddd defines a Julian date specifying the ending date of the period to place on the log archives file. Five blanks cause no records to be archived. The value TODAY can specify the current date. This parameter has no default. The nnnnnn defines a six-digit number, with leading zeros, representing the amount of memory available to the internal sort. If not used, it must be six blanks. The maximum available memory is the default.
When the SASSHIS5 program runs, a History Management report is written to the CNTLREPT DD. This report reflects the status of logged history or archive data which is stored as internal control data in the first record of the history file. The record provides information showing when the history data was written to the history file and when or if it was archived.
The SASSHIS6 program is used to manage the size of the history files, archive files, or both. SASSHIS6 deletes unwanted log records from the log history and log archives files while keeping the internal control data between the files synchronized. SASSHIS6 must run whenever the older log records are no longer needed for analysis. These records are permanently deleted from the system and prevent the log files from becoming too large. SASSHIS6 has a required PARM operand on the EXEC statement of PURGE= which identifies from and through dates defining the period (or range) of records to purge. You can DEMAND or schedule the execution of the SASSHIS6 program.
More Information:
For more information about the log data sets and the SASSHIS5 and SASSHIS6 processes, see the Systems Programming Guide.
|
Copyright © 2013 CA Technologies.
All rights reserved.
|
|