Monitoring zOS and OS390

Monitoring z/OS and OS/390

In addition to monitoring DB2, you must monitor the MVS system and its subsystems that communicate with DB2. Most MVS shops already support this type of monitoring. In this section, I outline the types of monitoring that should be established.

First, you should monitor memory use and paging system-wide for z/OS, for the DB2 address spaces, and for each DB2 allied agent address space (CICS, IMS/TM, and every TSO address space accessing DB2 ”both batch and online). A memory monitoring strategy should include guidelines for monitoring both CSA (common storage area) and ECSA (expanded common storage area). You can do so by using IBM's RMF (Resource Measurement Facility).

You should also monitor the CPU consumption for the DB2 address spaces. RMF can do this job.

You should also monitor the DASD space used by DB2 data. Underlying VSAM data sets used by DB2 for tablespaces and indexes must be properly placed on multiple data sets to avoid disk contention and increase the speed of I/O. They also must be monitored so that the number of data set extents is minimized, preferably with each data set having a single extent. This way, you can reduce seek time because multi-extent data sets rarely have their extents physically contiguous (thereby causing additional I/O overhead).

CICS and IMS/TM performance monitors should be available for shops that use these teleprocessing environments. IBM provides the CICS Monitoring Facility and CICSPARS for monitoring CICS performance, and the IMS/TM Monitor and IMSPARS for monitoring IMS/TM performance. The other leading vendors (BMC, Candle, and ASG) also supply monitors for CICS and IMS/TM.

Another monitoring task is to use a VTAM network monitor to analyze communication traffic. Finally, analysts can use other monitors to determine which statements in a single program are consuming which resources. This tool can be a valuable adjunct to RMF.