Many different activities are required to keep your system running smoothly, andto provide the best service on the basis of the available resources and workload requirements. The operator, the administrator, the system programmer, or the performance analyst will do these tasks. RMF is the tool that helps each of these people do the job effectively.
RMF consists of several components:
- Monitor I - Monitor II - Monitor III
- Spreadsheet Reporter
- Client/Server Enabling
- Sysplex Data Server
- Distributed Data Server for z/OS
- Distributed Data Server for AIX, Linux and Windows systems (RMF XP)
- RMF Performance Monitoring
These components work together in providing the capabilities you need for performance management:
- Gathering data
- Reporting data
- Accessing data across the sysplex
RMF gathers data using three monitors:
- short-term data collection with Monitor III
- snapshot monitoring with Monitor II
- long-term data gathering with Monitor I and Monitor III
The system operator starts all monitors as non-interactive (background) sessions with a variety of options that determine what type of data is collected and where it is stored. The data gathering functions run independently on each system, but each monitor can be started sysplex-wide by one operator command.
You can run data gathering on each z/OS system and use the RMF Sysplex Data Server to have all data available on the one system on which you run your performance management tasks.
Short-term data collection with Monitor III
The Monitor III gatherer session has a typical gathering cycle of one second, and consolidated records are written for a range which is typically set to 100 seconds. You can collect short-term data and continuously monitor the system status to solve performance problems. You get actual performance data (response times, execution velocity) on a very detailed level for later comparison with performance policy goals.
You can collect data that indicate how fast jobs or groups of jobs are running — this is called workflow or speed. You also get data that show how resource-intensive jobs are using the processor, the DASD devices, and the storage — the reports describe this under the term using.
There is also information about delays, which are important indicators of performance problems. This simplifies comparison of reports created from Monitor I and Monitor III data.
Snapshot monitoring with Monitor II
The scope of Monitor II data gathering is mainly related to single address spaces or resources, giving snapshots of the current status. You can collect data about address space activities and resource consumption, and about processor, DASD volume, and storage activities and utilization.
With Monitor II, it is also possible to monitor one specific job or volume continuously.
Long-term data gathering with Monitor I and Monitor III
Monitor I and Monitor III provide long-term data collection about system workload and resource utilization, and cover all hardware and software components of your system: processor, I/O device and storage activities and utilization, as well as resource consumption, activity and performance of groups of address spaces.
Data is gathered for a specific cycle time, and consolidated data records are written at a specific interval time. The default value for data gathering is one second and for data recording 30 minutes. You can select these options according to your requirements and change them whenever the need arises.
The SMF synchronization function ensures that records are written from all monitors in the sysplex for the same intervals.
RMF stores data in two types of record:
- All three monitors write SMF records (type 70 — type 79) if you define the appropriate SMF recording options.
- In addition, Monitor III writes VSAM records to in-storage buffers or into RMF-owned VSAM data sets.
All three monitors can create reports, and so does the Postprocessor.
Short-term interactive performance analysis with Monitor III
The Monitor III reporter runs in a TSO/E session under ISPF and provides sysplex or system performance reports by:
- Displaying your current system status in real-time mode
- Showing previously collected data that is still available in either in-storage buffers or preallocated VSAM data sets
Monitor III offers a wide spectrum of reports answering questions that arise during the various performance management tasks.
Cursor-sensitive control is one specific highlight of the Monitor III reporter you can use to navigate among different types of reports that all describe the system status at the same point in time from different perspectives. Once you have used it, you will never want to be without it — it helps you to get the report that points directly to the problems you need to solve.
All reporting is available within one TSO/E session, so there's no need to logon to different systems in the sysplex to get all performance data. All reports are available on one screen.
Snapshot reporting with Monitor II
Monitor II is a snapshot reporting tool for very fast information about how specific address spaces or system resources (processor, DASD volumes, storage) are performing. Monitor II has two modes for reporting on the performance of your system:
- Monitor II display session: You select the ISPF version in the RMF Performance Management menu, or you call the monitor with the TSO/E command RMFMON.
- Monitor II background session: You start a non-interactive session to create a report for printing.
Some reports offer continuous monitoring of single address spaces or DASD devices. You can get an one-line report each time you press ENTER, or you can request a periodically refreshed report.
Long-term overview reporting with the Postprocessor
Typically, you call the Postprocessor in a batch job, although running it in a TSO/E session is possible. You provide a set of options that define the scope of reporting and you get reports of various types with all the data you need for optimum running of your system.
The standard procedure is to allocate SMF data sets or SMF log streams with records from all monitors as input for the Postprocessor. A variation is to get reports on the RMF records that are available in the RMF data buffers of all systems in the sysplex while the Postprocessor is running. This data is automatically made available to the Postprocessor by calling the RMF Sysplex Data Server, a quick path to access performance data without having to go through dumping, sorting, and merging the SMF records.
The Postprocessor offers different types of report:
- Interval reports: they show the sysplex performance for each interval for which data has been gathered. Most single-system reports are also available as real-time reports from Monitor I.
- Duration reports: the data is summarized over longer periods of time with a maximum value of 100 hours.
- Summary, exception and overview reports: these Postprocessor capabilities let you create the reports you need to manage the performance of your system.
In addition, the Postprocessor can create Overview records which are the base for further spreadsheet processing on the workstation.
The Postprocessor can also generate a set of Postprocessor reports in XML format for display in a web browser.
Viewing reports on spreadsheets
The Spreadsheet Reporter is the function in RMF that assists you in converting Postprocessor listings and Overview records into spreadsheets. In addition, it provides sample spreadsheets to help you in presenting and analyzing performance data at a glance.
Monitoring on the workstation
IBM z/OS Management Facility (z/OSMF) is a web-browser based management console for z/OS. The z/OSMF Resource Monitoring plug-in allows cross-sysplex performance monitoring from a single point of control. From the z/OSMF task tree, you can select the following subtasks:
- The System Status task provides an enterprise-wide health check of all z/OS sysplexes.
- For further analysis, the Resource Monitoring task can graphically display RMF Monitor III metrics as well as AIX, Linux, or Windows metrics by means of customizable views.
RMF Performance Monitoring (RMF PM) gives you the capability to construct monitoring scenarios and use them whenever necessary. This is done on the Windows workstation, and the access to the current performance data of your z/OS systems is possible without the need to have a TSO/E session running.
Client/Server Enabling (RMFCS) uses the client/server concept to support performance management for z/OS systems without an active TSO/TCAS subsystem on the host.
You can access Monitor II and Monitor III reports with RMFCS by exploiting the ISPF Batch GUI feature. This way, RMFCS combines the advantages of a single point of control for z/OS performance management with a state-of-the-art user front end.
RMFCS supports event-driven monitoring. That is, predefined events on the MVS™ hosts can be configured to initiate performance monitoring. These events may be either specific system messages, or selected performance data counters that exceed predefined Monitor III exception thresholds.
Providing data for other applications
RMF Distributed Data Server (DDS): Applications that want to access sysplex-wide performance data, can retrieve their input from a single data server on one system in the sysplex, which gathers the data distributed on all systems in the sysplex. Therefore, this is called the Distributed Data Server (DDS).
The DDS offers an HTTP API which can access short-term data from the Monitor III as well as long-term data from the Postprocessor. An application program can
send an HTTP request for selected performance data to the DDS.
z/OS Common Information Model (CIM): This z/OS component allows access to RMF performance data from within systems management applications. These applications (called CIM clients) invoke the CIM server, which returns z/OS performance metrics collected by RMF Monitor III.
Resource monitoring of systems running AIX, Linux or Windows
The z/OS RMF Cross Platform (RMF XP) Distributed Data Server provides CIM-based performance data gatherers to monitor AIX on System p, Linux on System z, Linux on System x and Windows on System x. With RMF XP you can monitor operating systems which can run on an IBM zEnterprise System, including the zEnterprise Blade Center Extension (zBX).
You can exploit the RMF XP capabilities in the following ways:
- With the Resource Monitoring task of the IBM z/OS Management Facility (z/OSMF), performance metrics from connected AIX, Linux or Windows systems can be displayed in the same way and together with z/OS in heterogeneous customer environments.
- Exploiters of RMF XP can send an HTTP request to retrieve performance data from the endpoints running the AIX, Linux or Windows operating systems. Requests can be directed against an instance of the RMF XP core component called GPM4CIM, as soon as it is configured and running. GPM4CIM returns the requested data as a structured XML document.
Accessing data across the sysplex
The RMF Sysplex Data Server is a distributed RMF function. It is started as an identical copy on each system of the sysplex. Each copy of the data server communicates with all other copies in the sysplex. RMF uses this sysplex communication method to provide access to distributed RMF measurement data from any point in the sysplex.
The RMF Sysplex Data Server is always active when the RMF address space is running.
You can access all types of RMF and SMF data collected in the sysplex by using RMF Sysplex Data Server General-Use Programming Interface services. These are invoked as callable services by the RMF reporter sessions themselves or other applications, and can access:
- Monitor I, II and III SMF data
- Monitor III VSAM data
- SMF data of any other type
Sysplex data services for SMF data
RMF or other products store SMF data in a wrap-around buffer. You can choose to create a RMF Sysplex Data Server's SMF buffer when you start RMF. The size of the buffer and the types of SMF records stored in it can be specified as an RMF startup parameter. The RMF sysplex data services return SMF data when the RMF Sysplex Data Server's SMF buffer exists on at least one system in the sysplex, which need not be the system on which the calling program is running. The Data Server returns data only from systems in which data buffers have been created.
Sysplex data service for Monitor III data
You can access data collected by Monitor III data gatherer sessions using the RMF Monitor III Sysplex Data Retrieval Service. Any application program can specify the name of the system from which the Monitor III data is requested. Analogous to SMF data, Monitor III data can be returned from those systems where the Monitor III data gatherer session is active.
Sysplex data gathering service for Monitor II data
Your application program can use this service to create and retrieve Monitor II SMF records (type 79). You need not have a Monitor II background session running on the system from which you request the data. Note the difference between this service and the data service for SMF data, which collects only records created by active monitor sessions.