Allocate two sequential datasets (log datasets) to each job to collect logs.
-
Estimate the log dataset capacity.
Estimate the amount of log data. Allocate enough space to ensure that logs can be output until a backup of the log dataset is created in one sequential dataset.
-
Create log datasets.
For the log dataset, specify the amount of the initial allocation only. The specification of any additional allocation is ignored.
The following shows the format of the log dataset.
-
RECFM: VB
-
LRECL: 8196
-
BLKSIZE: 8200
-
DSORG: PS
-
DSNTYPE: BASIC
The following shows an example of creating log datasets in TSO/E.
ALLOC DS('USER01.RSLOG01') VOL(WKVL01) NEW SP(100) CYL DSORG(PS) RECFM(V,B) LRECL(8196) BLKSIZE(8200) DSNTYPE(BASIC)
ALLOC DS('USER01.RSLOG02') VOL(WKVL01) NEW SP(100) CYL DSORG(PS) RECFM(V,B) LRECL(8196) BLKSIZE(8200) DSNTYPE(BASIC)
-
Allocate the log datasets.
-
When using the START command, specify DD statement RSLOG01 and DD statement RSLOG02 in the cataloged procedure for startup.
-
When using JCL, specify DD statement RSLOG01 and DD statement RSLOG02 in the JCL.
The following shows specification examples for a cataloged procedure for startup and JCL.
//RSLOG01 DD DSN=USER01.RSLOG01,DISP=OLD,VOL=SER=WKVL01,UNIT=DASD
//RSLOG02 DD DSN=USER01.RSLOG02,DISP=OLD,VOL=SER=WKVL01,UNIT=DASD
Notes
Allocate a log dataset for each job. When the job terminates, another job can reuse that log dataset. In this case, logs are added to the end of the previous log output. Because the metric collection program has being running for a long time, allocate a different log dataset in another program.