Identifying and correcting performance problems
Your general strategy is to decide the origin of the problem being experienced by performing a set of tasks.
Complete the following tasks to decide on the origin of the problem:
- Isolate the problem into one of the following categories:
- High paging rate seems to be slowing some or all of the workload.
- High utilization of the processor by some subset of the total workload is causing dispatching problems for another (lower-priority) subset.
- I/O contention in the DASD subsystem is slowing some subset of the workload.
- Utilization of, or contention for, resources associated with the communications subsystem is slowing the transmission of input or output messages.
- Within IMS™, utilization of
physical resources (processor, I/O, storage) or contention for logical
resources (pools, regions, control blocks, latches) has a negative
impact on the performance of some or all of the transactions in one
of the following areas:
- Input or output message processing, including input queuing and Message Format Services (MFSs)
- Program scheduling and termination
- Program load and initialization
- Program execution
- Investigate further, if necessary, after you isolate the
problem area, to determine:
- The precise nature of the problem
- The principal offenders by transaction or other category
- The tuning action most likely to alleviate the problem
- Take the appropriate tuning action to prevent the problem from recurring.
- Examining paging rates
The operating system allocates real storage for the address spaces occupied by the control region or dependent regions. IMS might not obtain adequate machine cycles because of the excessive demand for real storage. Observe the paging rates and the frequency with which processing is delayed by a page fault. You can use the output from RMF™ II to examine the paging rates at peak IMS loads. - Detecting processor resource problems
In an environment with many jobs or subsystems operating concurrently, IMS competes for machine cycles. If those jobs or subsystems, such as an IMS batch execution or TSO, have a higher dispatching priority than the online IMS system, their workload has a marked effect on IMS performance. - Tuning to remove I/O resource contention
A performance problem can occasionally be traced to I/O contention. Tuning activities for I/O devices, DASD storage access, channel usage, or control unit contention should not be confined to system-related devices. For IMS, an important factor is the volume of I/O activity, because reduced I/O activity reduces contention. - Communication subsystem contention in DB/DC and DCCTL environments
Because of the great variety of transmission devices and types of lines, each suspected problem in the communication subsystem area must be separately examined. - IMS message processing in DB/DC and DCCTL environments
This topic identifies possible performance problems associated with message processing and the message queues on an IMS system. - Input queuing and scheduling/termination in DB/DC and DCCTL environments
Another set of performance problems is brought about by failure to achieve the performance objectives. The problem might be observed by input queue buildup caused by either increased workload or application program scheduling delays. - Program load and initialization
This topic describes dynamic and daily monitoring and detailed monitoring for program load and initialization. - Program execution times
You can obtain program execution time information with dynamic monitoring of IMS internal response times, daily monitoring of program execution, and detailed monitoring.
Parent topic: Tuning the system