Performance tools for Java in CICS Transaction Server
Version 3
Version 3
In this chapter, we introduce some important JVM performance monitoring and analysis tools on z/OS. These include:
•CICS Explorer®:
– IBM CICS Explorer: New Face of CICS. Integration point for CICS tooling with rich CICS views, data, and methods.
•CICS PA:
– CICS Performance Analyzer for z/OS is a reporting tool for analyzing and tuning the performance of your CICS systems.
•CICSPlex System Management:
– CICSPlex SM (commonly known as CPSM) is part of CICS Transaction Server and using it you can manage multiple CICS systems from a single control point.
•OMEGAMON XE for CICS on z/OS:
– OMEGAMON XE for CICS on z/OS provides the capability to monitor CICS Transaction Server (TS) environments.
10.1 CICS Explorer
The CICS Explorer is intended to act as the tooling integration point for the CICS runtime and to provide a rich set of CICS views, data, methods, and features, easily extensible by IBM, other vendors, and customers to deliver integrated solutions to key CICS users. The features offered by the CICS Explorer include:
•Common, intuitive, Eclipse-based environment for architects, developers, administrators, system programmers, and operators.
•Task-oriented views provide integrated access to broad range of data.
•Now packaged with Rational Developer for System z V7.5.
•Integration point for CICS TS, CICS Tools, Rational Tools, and others.
•Extensible by ISVs, SIs, and customers using the Software Development Kit bundled as part of the CICS Explorer.
•Dynamic resource relationships: See both uses and where used relationships for deployed CICS resources, for example, right-click a file to see all of the programs that use it (requires CICS IA® licence).
•Point-and-click navigation: Much easier to follow a sequence of resource relationships.
•Data filtering: Helps to reduce the volume of data displayed, so that you can identify the required relationships more easily.
•Dynamic perspectives: Resize, reorder, or sort columns. Move, tab, or resize views.
•Powerful online help: Browse, search, and print documentation with context-sensitive help and text search capability. The help is displayed in a Help view in the workbench, in a separate Help Contents window, or in an external browser window.
The CICS Explorer is not a replacement for the CPSM WUI. The CICSPlex SM WUI provides browser-based access to all CICSPlex SM function, and will continue to do so. The CICS Explorer provides an installed client that provides a rich, integrated user experience.
The CICS Explorer does not support editing of CICS resources at this time.
It is not our intention in this section to discuss all of the features and advantages of using TS Explorer. For complete information, see the CICS TS Explorer documentation.
10.1.1 System requirements
Table 10-1 shows the system requirements for the CICS Explorer. The CICS Explorer also requires a CICSPlex SM WUI server (CICS TS v3.1 or later) to connect to.
Table 10-1
|
Operating System
|
Software
|
Hardware
|
|
Linux, Windows
|
CICS Explorer runs under the control of, or in conjunction with, the following programs and their subsequent releases, or their equivalents:
•Eclipse V3.3 (included)
•Windows: Release to be determined
•Linux: Release to be determined
|
CICS Explorer runs on any hardware configuration that is supported by the licensed programs specified below.
|
|
z/OS
|
Required licensed programs:
•IBM z/OS, Version 1.7 or later
•CICS Transaction Server for z/OS V3.1 V3.2 (5655-M15)
|
CICS Explorer runs on any hardware configuration supported by the licensed programs specified below.
|
TS Explorer scene layout
A connected TS Explorer view looks similar to Figure 10-1. The window layout can vary a little, depending on what you are doing in the Explorer, but you can typically view it as three panels: left, top right, bottom right, and a status bar and the menus.
Figure 10-1 Typical Explorer view
The left panel shows the CPSM CICSplexes and the CICS regions. They contain what we can view through the server that we are connected to. Because we are connected to a CPSM WUI server, we benefit from the grouping of CICS regions into logical scopes, as provided by CPSM. Selecting a region, CICSplex or other logical scope in this view changes the scope of the other views, so that they show the information for that scope.
The top-right panel can show a variety of different views. Each view shows different aspects of the region or regions in the currently selected scope. There are a variety of views that are available that correspond to most of the resources that exist within a CICS region. There are several views that relate to Java in CICS.
The bottom-left panel also can show a variety of different views. The data displayed here often is dependent on what you select in the top-right panel. The properties tab, shown here, shows more information about the currently selected resource.
Many of the views have context menus (brought up by selecting the right-mouse button) that allow further information to be displayed about a given resource. If you have additional products (such as CICS Interdependency Analyzer or CICS Configuration Manager) installed then using these menu options, you can view additional information and apply changes to your CICS regions through those tools.
Do not overlook the status bar because it includes information, such as, the fact that we are connected to a server and provides name that we gave to the connection to the server.
You can customize the view by resizing, moving, or even removing panels. If you move these panels around, there is a menu option, Window ∅ Reset Perspective, that restores the panels to their original positions and brings back any panels that you removed from the view.
One of the more interesting menus is the ‘operations’ menu, which allows you to choose which of the CICS resources you want to see more information about. There is a selection of resource views that you can select, and we briefly review the Java related resources here. Under the Operations ∅ Java menu, the following views are available:
•JVM Class Caches
•JVM Pools
•JVM Profiles
•JVM Status
We describe each of these views below.
JVM class caches
Class cache is used for JVM shared classes. This view, Figure 10-2, gives the status of the class cache and the number of JVMs being phased out.
Figure 10-2 JVM Class Cache view
JVM pool
Each JVM runs on an MVS TCB, which is allocated from a pool of J8- and J9-mode open TCBs, which the CICS manages in the CICS address space.
This view, Figure 10-3 on page 237, gives you the status of the pool, the number of JVMs that are removed from the pool when they finished executing, and the number of pre-initialized JVMs.
If a JVM is not used by any application during the period of time that is specified in the IDLE_TIMEOUT option in its JVM profile, it becomes eligible for automatic termination.
Figure 10-3 JVM Pool view
JVM profiles
Each JVM requires a profile that contains the properties that are needed for a particular JVM, for example, you can specify different JVM Profiles according to the type of program that is going to execute in the JVM. Figure 10-4 shows the JVM profile view.
Figure 10-4 JVM profile view
JVM status
This view, shown in Figure 10-5, displays the status of every JVM in the current scope.
Figure 10-5 JVM status view
10.2 CICS PA overview
CICS PA is a reporting tool that provides information about the performance of your CICS systems and applications to help you tune, manage, and plan your CICS systems effectively.
CICS PA is not an online monitoring tool. It produces reports and extracts using data that your system collects in MVS System Management Facility (SMF) data sets:
•CICS Monitoring Facility (CMF) performance class, exception class, and transaction resource class data in SMF 110 records
•CICS Transaction Server statistics data in SMF 110 records
•CICS Transaction Gateway statistics data in SMF 111 records
•System Logger data in SMF 88 records
•DB2 accounting data in SMF 101 records
•WebSphere MQ accounting data in SMF 116 records
•IBM Tivoli® OMEGAMON XE for CICS on z/OS (OMEGAMON XE for CICS) data in SMF 112 records, containing transaction data for Adabas, CA-Datacom, CA-IDMS, and Supra database management systems
CICS PA can help:
•System Programmers to track overall CICS system performance and evaluate the results of their system tuning efforts
•Application Programmers to analyze the performance of their applications and the resources they use
•Database Administrators to analyze usage and performance of database systems, such as IMS and DB2
•WMQ Administrators to analyze usage and performance of their WebSphere MQ messaging systems
•Managers to ensure that transactions are meeting their required Service Levels and measure trends to help plan future requirements and strategies
CICS PA reports all aspects of CICS system activity and resource usage, including:
•Transaction response time
•CICS system resource usage
•Cross-system performance, including multi-region operation (MRO) and advanced program-to-program communication (APPC)
•CICS Business Transaction Services (BTS)
•CICS Web Support
•External subsystems, including DB2, IMS, and WebSphere MQ
•System Logger performance
•Exception events that cause performance degradation
•Transaction file and temporary storage usage
Rather than keeping large SMF data sets for reporting purposes, you can use CICS PA to load selected SMF records into a CICS PA historical database (HDB), optionally summarizing the records according to the time intervals that you require for reporting (such as hourly or daily). You can then use CICS PA to produce reports from the HDB instead of the SMF data sets. Loading selected and summarized SMF data into an HDB, you can accumulate the performance data that you want at the level of detail that you need for reporting over long periods, without requiring large amounts of storage or processing time.
In addition to producing formatted reports from SMF data sets or HDBs, CICS PA can extract data to DB2 tables or comma-separated value (CSV) text files. You can then develop your own custom reports using DB2 SQL queries or download CSV files to your PC, where you can view and manipulate the data using PC-based spreadsheet applications, such as Microsoft Excel.
CICS PA provides both an interactive ISPF dialog interface and a batch command interface. You can use either of these interfaces to request your reports and extracts. The ISPF dialog interface uses your interactive input to prepare JCL for the batch command interface. If you prefer to work directly with a command interface rather than an interactive interface, then you can use the ISPF dialog interface to prepare JCL that you can save and use as a starting point, and then edit the JCL later without using the ISPF dialog.
Figure 10-6 shows the SMF record types that CICS PA can read, and the output formats that it can produce.
Figure 10-6 Overview of CICS PA inputs and outputs
Figure 10-7 on page 240 shows the primary option menu of the CICS PA ISPF dialog.
Figure 10-7 CICS PA primary option menu
Primary menu option 2, Report Sets, displays a list of the formatted reports that you can select, which are organized by category. Figure 10-8 on page 241 shows a sample report set. The Active column indicates the reports that this particular report set contains. You can generate and submit JCL to run reports individually in separate batch jobs, or you can submit an entire report set in a single batch job and perform a single pass over the input data.
Figure 10-8 CICS PA report set
Measuring JVM performance using CICS PA
Using CICS PA, you can measure JVM performance too. After the SMF data sets were switched, we ran the Take-up facility on the SMF file. In CICS Performance Analyzer Primary Option Menu, we chose option 1 to go to the Systems Definitions window. On the Systems Definitions window, we chose option 4 for the Take-up function. On the resulting window, we entered the SMF records data set name and submitted the batch job. Figure 10-9 on page 242 shows output from the take-up job.
|
V2R1M0 03:59:48 11/05/2008 CICS Performance Analyzer Page 1
Personal Systems Take-up from SMF
_________________________________
CPA2012I Processing started for SMF file SMFIN001
CPA2017I SMF records for System SC66 start at 11/05/2008 07:01:00.05
CPA2014I CMF record for CICS system found, APPLID=A6POC3C1 Release=6.5.0
CPA2014I CMF record for CICS system found, APPLID=A6POS3C2 Release=6.5.0
CPA2016I MVS System Logger record found, System=SC66LOGR
CPA2013I Processing ended for SMF file SMFIN001 - 3 system(s) found
CPA2000I Personal Systems Take-up processing has completed, RC=0
|
Figure 10-9 Take-up output
Before running the first report, we chose a Report Form. We chose to use the CICS PA provided sample forms that are related to JVM. To bring these to our forms data set, on the CICS PA Primary Option Menu, we chose option 3 to go to the Report Forms panel. On this panel, we moved the cursor to the action bar under the Samples option and pressed Enter. We selected option 1 to populate the Report Forms data set with the sample forms, as shown in Figure 10-10.
|
Sample Report Forms Row 77 to 89 of 156
Command ===> Scroll ===> PAGE
Select one or more Sample Report Forms then press EXIT.
Name Type Description
JVMLST LIST Java Virtual Machine Analysis
JVMSUM SUMMARY Java Virtual Machine Analysis
OMDLMLST LIST OMEGAMON Database Limit Warnings
OMOEMLST LIST OMEGAMON Third Party Support
OMOEMSUM SUMMARY OMEGAMON Third Party Support
OMRLMLST LIST OMEGAMON Resource Limit Warnings
PCLST LIST Program Request Activity
PCSUM SUMMARY Program Request Activity
PC3LST LIST Program Request Channel Activity
PC3SUM SUMMARY Program Request Channel Activity
PGAPLSUM SUMMARY Transactions by Application Prog
PGDPLSUM SUMMARY DPL Program Usage by Connection
PGUSESUM SUMMARY Transactions by Initial Program
|
Figure 10-10 Sample Report Forms pop-up panel
We scrolled down the list and selected the JVMLST and JVMSUM sample forms. We pressed F3 to return to our private Report Forms panel where we saw these forms included in the list.
We first produced the LIST report. We returned back to the Primary Option Menu, where we selected 2 to go the Report Set panel. On the Report Set panel, we entered NEW JVML to create a new Report Set. On the EDIT Report Set panel, we entered S next to the List option in the category Performance Reports. On the Performance List Report panel, Figure 10-11, we entered the APPLID and image name as generated by the take-up, specified to use the JVMLST Report Form, and entered a report title.
|
File Systems Options Help
-----------------------------------------------------------------------------------------------
JVML - Performance List Report
Command ===>
System Selection: Report Output:
APPLID . . A6POC3C1 + DDname . . . . . . . . . LIST0001
Image . . SC66 + Print Lines per Page . . (1-255)
Group . . +
Report Format:
Form . . . JVMLST +
Title . . LIST report of JVM related fields
Selection Criteria:
Performance
|
Figure 10-11 Performance List Report selections
We returned to the Edit Report Set panel where we ran the List report. We entered the time interval for our report, as shown in Figure 10-12 on page 244, and then submitted the report.
|
File Systems Options Help
---------------------------------------------------------------------------------------------
Run Report Set JVML
Command ===>
Specify run options then press Enter to continue submit.
System Selection:
CICS APPLID . . A6POC3C1 + Image . . + Group . . +
DB2 SSID . . . + Image . . + Group . . +
MQ SSID . . . . + Image . . + Group . . +
Logger . . . . + Image . . + Group . . +
Override System Selections specified in Report Set
sssss Report Interval ssssss
Missing SMF Files Option: YYYY/MM/DD HH:MM:SS.TH
1 1. Issue error message From 2008/11/05 01:00:00.00
2. Leave DSN unresolved in JCL To 2008/11/05 03:00:00.00
3. Disregard offending reports
Enter "/" to select option
/ Edit JCL before submit
|
Figure 10-12 Time interval selection for JVML report
Figure 10-13 shows the Performance List report.
V2R1M0 CICS Performance Analyzer
Performance List
__________________________________________________
LIST0001 Printed at 22:09:39 11/10/2008 Data from 01:59:51 11/05/2008 APPLID A6POC3C1 Page 1
LIST report of JVM related fields
Tran Userid TaskNo Stop Response Dispatch User CPU J8 CPU J9 CPU JVM Elap JVMITime JVM Meth JVMRTime JVM Susp
Time Time Time Time Time Time Time Time Time Time Time
CWBA CICSUSER 76078 02:33:49.958 .0015 .0014 .0013 .0000 .0011 .0011 .0000 .0011 .0000 .0001
CWBA CICSUSER 76080 02:33:51.482 .0027 .0020 .0017 .0000 .0014 .0020 .0000 .0020 .0000 .0004
CWBA CICSUSER 76082 02:33:52.136 .0016 .0014 .0014 .0000 .0011 .0011 .0000 .0011 .0000 .0001
CWBA CICSUSER 76084 02:33:55.066 .0064 .0062 .0060 .0000 .0057 .0059 .0000 .0059 .0000 .0001
CWBA CICSUSER 76086 02:33:55.819 .0017 .0015 .0015 .0000 .0011 .0012 .0000 .0012 .0000 .0001
CWBA CICSUSER 76088 02:34:04.636 .0134 .0118 .0066 .0000 .0044 .0111 .0000 .0110 .0000 .0008
CWBA CICSUSER 76091 02:34:05.245 .0017 .0015 .0014 .0000 .0011 .0012 .0000 .0012 .0000 .0001
Figure 10-13 The Performance List report
Conclusion
CICS Performance Analyzer for z/OS is a good reporting tool that helps you tune, manage, and plan your CICS systems in an efficient way. At the same time, CICS Performance Analyzer for z/OS supports JVM as well. Using CICS PA, you can analyze utilization and response time including JVM, KEY8 CPU, J8 CPU, and so on.
You can download CICS PA SupportPac CP12 from the Web at no charge:
10.3 CICSPlex System Management
CICSPlex SM (commonly known as CPSM) is part of CICS Transaction Server and using it you can manage multiple CICS systems from a single control point. Enterprises who use CICSPlex SM range from those who have 10 CICS systems to those that run hundreds of CICS systems.
It is not our intention in this section to discuss all of the features and advantages of using CICSPlex SM. For complete information, access the CICS Infocenter that is applicable to the level of CICS Transaction Server that you installed and refer to the book entitled CICSPlex SM Concepts and Planning.
Using CICSPlex SM you can:
•Manage resources using Business Application Services (BAS)
BAS is an alternative to resource definition online (RDO) that enables you to manage CICS resources in terms of the business functions they belong to rather than their location in the CICSPlex.
Using BAS has a number of advantages over RDO:
Using BAS has a number of advantages over RDO:
– The process is similar to RDO with a choice of interfaces
– Logical scoping allowing you to handle resources in terms of business application
– A common repository for all the resources in a CICSPlex
– Reduced number of resource definitions leading to a consistency in definitions
– You can add resources to a region simply by adding it to the groups where the application executes
– You can direct CICSPlex SM commands to a scope that matches the application instead of an arbitrary group of regions
•Manage Workloads
CICSPlex SM workload management optimizes processor capacity in your enterprise by dynamically routing transactions and programs to which ever CICS region is the most appropriate at the time, taking into account any transaction affinities that might exist.
To achieve true workload balancing, design and write applications so that transaction affinities are either removed or kept to an absolute minimum.
There is a big difference between workload distribution and workload balancing. Workload distribution is distributing the work around the enterprise without necessarily taking into consideration how busy the target regions are and is often implemented using a “round robin” approach where target CICS regions are selected sequentially. Workload management can achieve workload balancing across the enterprise, provided the workload is actually high enough to be managed. This distinction is often a source of confusion and needs to be understood.
There are two algorithms that CICSPlex SM uses to achieve workload balancing:
– The Queue algorithm
CICSPlex SM sends work to the region that:
• Has the shortest queue of work waiting to be processed
• Is least likely to abend or meet conditions, such as SYSDUMP or TRANDUMP
– The Goal algorithm
CICSPlex SM sends work to the region that:
• Meets the average response time goals set for it using the Workload Manager component of z/OS
• Is least likely to abend or meet conditions such, as SYSDUMP or TRANDUMP
When using the Goal algorithm you need to be aware that if one particular CICS region is meeting the goals set for the transaction the likelihood is that CICS Plex SM will continue sending work to this region until a condition occurs or the response time goal is not met.
•Exception Reporting using real-time analysis (RTA)
RTA provides automatic, external notification of conditions in which you might have expressed an interest, for example, if a FILE must always be ENABLED, the RTA component sends out warning messages whenever it finds that the file is not ENABLED.
The alerts can be sent either to the console or NetView® or both:
– System Availability Monitoring (SAM)
This function monitors CICS systems during their planned hours of availability. If any of a set of predefined conditions occurs while the systems are being monitored, CICSPlex SM sends out external notifications at the start of the condition and again when it is resolved.
– MAS resource monitoring (MRM)
You can use this to monitors the status of any specific or generic CICS resource and be informed when the state changes from a specified state, for example, when a critical FILE is CLOSED.
– Analysis point monitoring (APM)
When a resource is monitored using MRM in a number of regions there is a possibility that multiple notifications are sent when the state changes from the expected state. This function can be used to combine these messages into a single notification.
•Collecting statistics using CICSPlex SM monitoring
CICSPlex SM monitoring supports the collection of performance-related data at user-defined intervals within a set of CICS systems and can be used instead of the CICS Monitoring Facility (CMF)
The CICSPlex SM Web User Interface
The CICSPlex SM Web User Interface (WUI) is a customizable, platform-independent interface for your Web browser.
It is supplied with a set of linked menus and views to facilitate all of your system management tasks. You can customize the WUI to reflect your business procedures and the needs of individual users.
The CICSPlex SM Web User Interface uses a frame-based interface.
•The Navigation Frame appears on the left of the display and contains items that you can use to display a menu or to view or perform an action.
•The Work Frame is the area where data is presented to you for interaction. Using this frame, you can set the Refresh rate and request a refresh by clicking the Refresh button.
•The Assistance Frame contains the product name, an icon linking to the Web User Interface Help, the IBM logo, and the Go back to start, Go back to last menu, and Go back icons.
For full details about the CICSPlex SM Web User Interface, read the CICSPLEX SM Web User Interface Guide applicable to your installation.
The initial view
After you successfully logon to the Web User Interface, you are presented with this initial view of your CICSPlex.
The Work Frame names the CMAS context, the Context, the Scope, and the views that are available to you and all of the links in the Navigation frame are closed.
Most views have the option to refresh the window by either setting a value for automatic refresh or a button for instant refresh.
Figure 10-14 on page 248 illustrates the Main menu.
Figure 10-14 The CICSPlex SM supplied Main Menu
The Enterprise Java link
The Enterprise Java link is a clickable link in the navigation frame. Clicking it expands to give you links to the following views:
•Corbaservers
•CICS-deployed JAR files
•Enterprise Beans in CorbaServer
•Enterprise Beans in CICS-Deployed JAR files
•JVM Pool
•JVM Profile
•JVM Status
•JVM class cache
Figure 10-15 on page 249 shows the Enterprise Java link view.
Figure 10-15 The Expanded Enterprise Java link
Corbaservers
The Corbaservers view, Figure 10-16 on page 250, shows all of the Corba servers that are defined in your CICSPlex.
The fields “CorbaServer name” and “TCP/IP host address” are links to views that give you detailed information about each resource.
There are four Action buttons available that you can use to manage the CorbaServer:
•Perform a CorbaServer scan to pick up any updated JARs in the zFS directory that is specified on the DJARDIR parameter of the CorbaServer definition
•Publish all Enterprise Java beans that are associated with the CorbaServer to the name server that is specified in the JVM properties file
•Retract all Enterprise Java beans that are associated with the CorbaServer from the name server that is specified in the JVM properties file
•Discard the CorbaServer resource
Figure 10-16 CorbaServers
CICS-deployed JAR files
These JAR files, shown in Figure 10-17 on page 251, are installed from the DJARDIR parameter that is specified on the CORBASERVER definition and are automatically installed by CICS after the CORBASERVER is installed.
The fields CICS-deployed Jar file, CorbaServer name, and Hierarchical file name (HFS) path are links to subsequent views that give detailed information about each resource.
There are three Action buttons available that allow you to manage the CICS-deployed JAR file:
•Publish all Enterprise Java beans that are associated with the CICS-deployed JAR file to the name server that is specified in the JVM properties file
•Retract all Enterprise Java beans that are associated with the CICS-deployed JAR file from the name server that is specified in the JVM properties file
•Discard the CICS-deployed JAR resource
Figure 10-17 CICS-deployed JAR files
Enterprise beans in CorbaServer
These JAR files, Figure 10-18 on page 252, are installed from the DJARDIR parameter that is specified on the CORBASERVER definition and are automatically installed by CICS after the CORBASERVER is installed.
The fields CorbaServer name, Enterprise Bean name, and CICS-deployed JAR file are links to views that give detailed information about each resource.
Figure 10-18 Enterprise beans in CorbaServer
Enterprise beans in CICS-deployed JAR file
This is a slightly different view to Enterprise Beans in CorbaServer. Here, the view lists each Enterprise bean in the CICS-deployed JAR file and the associated CorbaServer.
The fields CICS-deployed JAR file, Enterprise bean name, and CorbaServer name are links to views that give detailed information about each resource.
Figure 10-19 on page 253 shows the Enterprise beans in CICS-deployed JAR file window.
Figure 10-19 Enterprise Beans in CICS-deployed JAR file
JVM pool
Each JVM runs on an MVS TCB, which is allocated from a pool of J8- and J9-mode open TCBs, managed by CICS in the CICS address space.
The JVM pool view, Figure 10-20 on page 254, gives you the status of the pool, the number of JVMs to be removed from the pool when they finish executing, and the number of pre-initialized JVMs.
If an application does not use JVM during the period of time that is specified in the IDLE_TIMEOUT option in its JVM profile, it becomes eligible for automatic termination.
There are seven action buttons available that you can use to manage the JVM:
•Set Attributes to Enable or Disable the JVM Pool
•Enable: Enables the JVM Pool
•Disable: Disables the JVM Pool, which prevents new requests from being serviced from the pool. Currently, executing JVMs are allowed to terminate normally
•Phaseout: Marks all the JVMs for deletion and allows work that is currently running to complete
•Purge: Purges all tasks using the JVMs, and then terminates all JVMs using the pool
•Forcepurge: Force purges all tasks using the JVMs and then terminates all JVMs that are using the pool
Figure 10-20 JVM pool
JVM Profile
Each JVM requires a profile that contains the properties that are needed for a particular JVM, for example, you can specify different JVM profiles according to the type of program that is going to execute in the JVM.
The field “Name as used in a program definition” is a link to a view for detailed information.
Figure 10-21 on page 255 shows the JVM Profile view.
Figure 10-21 JVM Profile
JVM status
The JVM status view, Figure 10-22 on page 256, displays the status of every JVM in the CICSPlex. The Java Virtual Machine field is a link to a view that gives detailed information about the JVM, for example, the JVMProfile that is used to initialize the JVM.
Figure 10-22 JVM status
JVM class cache
Class cache is used for JVM shared classes. The JVM class cache view gives the status of the class cache and the number of JVMs that are phased out.
The “CICS system name” is a link to a view that gives more information about the class cache, such as time started, size of the class cache, and amount of free space in the class cache.
There are five action buttons available that you can use to manage the JVM class cache:
•Phaseout: Marks all of the JVMs that are using the shared class cache for deletion, which allows work that is currently running to complete.The shared class cache is deleted when all JVMs that are using it are terminated
•Purge: Purges all JVMs using the shared class cache. The shared class cache is deleted when all JVMs using it are terminated
•Forcepurge: Force purges all JVMs using the shared class cache. The shared class cache is deleted when all JVMs using it are terminated
•Start: Creates a new class cache when the status is STOPPED
•Reload: Creates a new class cache when the status is STARTED
Figure 10-23 JVM class cache
10.4 OMEGAMON XE for CICS on z/OS
OMEGAMON XE for CICS on z/OS is the component of IBM Tivoli Monitoring that provides the capability to monitor CICS Transaction Server (TS) environments. It is installed as a monitoring agent in the IBM Tivoli Monitoring (ITM) framework. It provides workspaces and situations that provide alerts when components of the CICS TS environment are not meeting the expected performance parameters.
As a component of ITM, OMEGAMON XE for CICS also provides the capability to combine CICS monitoring with that of other monitored components of the enterprise. It uses Dynamic Workspace Linking (DWL) to allow a user to switch between OGEMAON XE for CICS and other OMEGAMONs, such as OMEGAMON XE for DB2 on z/OS, which gives users the capability to follow transactions to identify the route of a problem.
OMEGAMON XE for CICS on z/OS also provides a 3270 interface that provides the capability to perform in depth diagnosis of a CICS TS system.
In this section, we explain the major components of OMEGAMON XE for CICS and describe some of the major features. This is not a comprehensive description, but you can use it to gain an understanding of the product.
OMEGAMON XE for CICS components
For OMEGAMON XE for CICS to provide information relating to CICS Transaction Server, the following components must be configured on the z/OS image where the monitored CICS regions reside:
•An OMEGAMON XE for CICS Monitoring Agent
•An OMEGAMON II for CICS Menu System
•An OMEGAMON II for CICS CUA interface (optional)
•OMEGAMON for CICS component in monitored CICS regions
•Configuration of OMEGAMON II for CICS
In the next sections, we discuss the function and specific requirements of each of these components in more detail. For the purposes of this chapter, we assume that the ITM framework is installed and configured.
OMEGAMON XE for CICS Monitoring Agent
The OMEGAMON XE for CICS monitoring agent is the component that is responsible for responding to queries from the ITM framework for data relating to the CICS TS systems that are currently monitored. The agent obtains information by interfacing with OMEGAMON code that runs in the CICS address space. It also communicates directly with the Menu System started task (KOCCI) to query data that it maintains. The agent also has a Workload Manager (WLM) component that generates summaries of transaction performance against specified goals.
The monitoring agent provides the data that is required to populate the CICS workspaces on the TEP. Figure 10-24 shows an example of a workspace that is provided with the OMEGAMON XE for CICS on z/OS product.
Figure 10-24 OMEGAMON XE for CICS region overview
An agent must be deployed on each LPAR where CICS TS regions run. It is recommended that the OMEGAMON CICS agent run its own address space.
The OMEGAMON XE for CICS agent is configured using the Installation and Configuration Assistance Tool that is provided with ITM.
Figure 10-25 shows the configuration page for the OMEGAMON XE for CICS agent.
|
------ CONFIGURE IBM TIVOLI OMEGAMON XE FOR CICS ON Z/OS / RTE: SC66 ----------
OPTION ===>
Last selected
Perform the appropriate configuration steps in order: Date Time
I Configuration information (What's New)
1 Register with local TEMS (required if the Agent 07/09/24 09:40
will connect to the TEMS in this RTE.)
2 Specify configuration parameters 08/08/28 16:56
Agent address space configuration:
3 Specify Agent address space parameters 08/08/28 16:56
4 Create runtime members 08/08/28 16:58
5 Configure persistent datastore (in Agent) 08/08/28 16:58
6 Complete the configuration 07/08/22 15:22
Note: This Agent is running in its own Agent address space.
F1=Help F3=Back F5=Advanced
|
Figure 10-25 Configuration page for OMEGAMON XE for CICS on z/OS
Notes about the OMEGAMON configuration page options:
•Option 1, Register with local TEMS: Adds application support files to the TEMS. Here is a summary of when you are required to select this option:
– If the agent you are configuring reports to a TEMS that is configured in the same (that is, local) RTE.
– If your hub TEMS is running on z/OS but the agent reports to a remote TEMS, register the monitoring agent with both the remote TEMS it directly reports to and the hub TEMS.
– If the agent reports to the hub TEMS directly, you only need to register it one time. If the hub TEMS is configured in a separate RTE from the agent, select that RTE for configuration of the OMEGAMON XE agent, but only select option 1 (Register with local TEMS). None of the other configuration items need to be completed unless you are also going to define an OMEGAMON XE agent in the same RTE as the Hub TEMS.
– If the agent reports directly to the Hub TEMS, but this TEMS is configured in a different RTE from the agent, you must select that RTE for configuration and then only select option 1 (Register with local TEMS) for this configuration. None of the other selected agent configuration items that follow should be performed.
In our configuration, where we want to report directly to a remote TEMS in the same RTE as the monitoring agent and our hub TEMS is installed on another operating system, we must register our monitoring agent just one time with the remote TEMS.
•In option 2, Specify configuration parameters: Shown in Figure 10-26 on page 260, we specify the configuration options that are unique to OMEGAMON XE for CICS on z/OS.
|
----------------------- SPECIFY CONFIGURATION PARAMETERS ----------------------
OPTION ===>
Complete the items on this panel.
WLM block allocation ==> 236 (10-524287 blocks)
WLM collection interval ==> TEP (TEP or 1 minute)
Enter=Next F1=Help F3=Back F5=Advanced
|
Figure 10-26 OMEGAMON XE for CICS on z/OS product specific parameters
The WLM block allocation specifies the number of 4096 byte blocks that are allocated to the WLM data space. This data space stores information while summaries are created. Although the usage of this storage is stable for a given workload, it is not easy to predict. It is recommended that the number start at 236. If the message KCP0244 is produced by the agent, this indicates that the value must be increased.
The WLM collection interval controls the time span of Service Level Analysis displays at the TEP. OMEGAMON WLM accumulates data after a minute in two sets of records, the "5 minute records" and the "interval records". The "5 minute records" are displayed at the TEP. The "interval records" are used for history (Persistent Data Store).
Calculations for values, such as average response time, are performed every five minutes and at the end of the history interval.
Specifying a value of 1 transforms the "5minute records" to "1minute records". Starting WLM with a frequency interval value of 1 results in the TEP displaying Service Level data collected at the last "1minute" interval.
Specifying a value of TEP results in the TEP display of the "5minute records" using the values specified via TEP.
The acceptable values for the INTERVAL keyword are 1 or TEP. You can specify values other than 1 using the SLA CICS view of the TEP.
Options 3, 4, and 5 are common to all of the OMEGAMON XE monitoring products. These steps are required to be completed if you are configuring the agent to run in its own address space.
•Option 3, Specify Agent address space parameters: Where you specify the communication values for reporting to the hub and the “self-describing” values for configuring up your agent address space, such as security, locale, and so on.
•In option 4, Create runtime members: A batch job is generated that updates the user parameter libraries with the values you specified as part of configuration options 2 and 3. After submitting this job check, the return code produced no errors.
•Option 5, Configure persistent data store (in Agent): Allocates the persistent data store files and starts up parameters as part of the agent address space. During the persistent data store configuration, you are asked to enter JCL jobcard information. Note that this jobcard information is shared by all persistent data store runtime JCL generated for this RTE.
If you have a TEMS defined in this RTE and want to run the agent in the local TEMS address space, you must select F5, and follow the advanced configuration options. This is not the recommended agent configuration, so we do not discuss it in this book.
•Option 6, Complete the configuration: Guides you through some manual configuration steps that must be performed outside of the Configuration Tool, such as copying the agent started task to your PROCLIB and ensuring libraries are APF authorized. These steps are required to be completed prior to testing your configuration.
OMEGAMON II for CICS Menu System
The OMEGEMAON II for CICS menu system consists of a started task, known as the common interface, and also referred to as the KOCCI. It provides a 3270 interface to allow monitoring of the CICS regions on the same LPAR. It provides an anchor to the OMEGAMON for CICS components that run in a CICS region. It is responsible for loading of the Global Data areas that control which features of OMEGAMON II are to be active. It also provides an environment for the subtasks that are required to provide Bottleneck Analysis, Response Time Analysis, and Transaction History collection.
The 3270 interface that the KOCCI provides gives the user a highly efficient mechanism to view specific information relating to a CICS region. The menu system allows you to see resource detail information and detailed information regarding the memory usage and DASD performance of the devices that the CICS region currently uses. The Menu system also allows authorized uses to view and modify the storage in the CICS region.
Figure 10-27 shows the primary panel that the menu system provides.
|
________________ ZMENU VTM A6POC3C1 V560./C SC66 09/21/08 12:08:24
> PF1 Help/News/Index PF3 Exit PF18 Color PA2 REGION STATUS
===============================================================================
> OMEGAMON II FOR CICS PERFORMANCE MONITOR SYSTEM
> Enter a selection letter on the top line.
> W REGIONS ............ List CICS regions, switch monitoring
> V OVERVIEW ........... Performance overview
> R RESPONSE TIME ...... CICS and end-to-end response time monitoring
> E EXCEPTIONS ......... Current system problems and problems this session
> T TASKS .............. Task analysis
> H HISTORY ............ Historical, traced transaction viewing and selection
> B BOTTLENECKS ........ Resource contention (bottlenecks, impacts, enqueues)
> S STORAGE ............ Storage summary, violations, DSA, EDSA, PAM, subpools
> F FILES .............. CICS datasets, VSAM, LSR, string and buffer waits
> D DATABASES .......... DB2, DLI and MQ
> C CICS ............... CICS tables and control blocks
> M MVS ................ Operating system control blocks for CICS
> I I/O ................ DASD performance
> U UTILITIES .......... Utility functions and displays
> O CONTROL ............ Control options (response time, history, contention,
> database collectors, shutdown, SMF, trace)
> G GROUPS ............. Group definitions
> P PROFILE ............ Profile options and maintenance
===============================================================================
|
Figure 10-27 OMEGAMON II menu system interface
The menu system can only monitor CICS regions that are running on the same LPAR. For that reason, a KOCCI must be configured on each LPAR where CICS regions that are to be monitored can run.
OMEGAMON II for CICS CUA interface (optional)
The OMEGAMON II for CICS CUA interface provides a means of viewing the data provided by the Menu system that conforms to the Common User Access (CUA) standard.
The interface provides an easy point and shoot type access to the monitoring data. An example of the CUA Region Overview panel is provided in Figure 10-28 on page 262.
Figure 10-28 CUA region overview
The CUA interface is not a required part of the OMEGAMON CICS configuration. The menu system and OMEGAMON XE for CICS provide full functionality, regardless of the status of the CUA interface.
OMEGAMON for CICS component in monitored CICS regions
For OMEGAMON CICS to provide full functionality, you must make certain changes to the CICS regions that are to be monitored. OMEGAMON code needs to be installed and run in each CICS region to allow OMEGAMON to enable certain CICS Global User Exits (GLUEs) to allow the collection of data that is required for features, such as Transaction History, Service Level Analysis, Response Time Analysis, and some resource monitoring.
OMEGAMON II for CICS requires that a Program and a Transaction be defined, as shown in Figure 10-29.
|
DEFINE PROGRAM(KOCOME00) GROUP(OMEGAMON) CONCURRENCY(THREADSAFE)
EXECKEY(CICS)
DEFINE TRANSACTION(OMEG) PROGRAM(KOCOME00) GROUP(OMEGAMON)
TASKDATAKEY(CICS)
|
Figure 10-29 RDO definitions for OMEGAMON CICS
CICS PLT changes
For OMEGAMON CICS to be configured to start and stop automatically, the CICS PLTs must be updated. Here are the PLT additions for installation:
DFHPLT TYPE=ENTRY,PROGRAM=KOCOME00
Specify the PLT additions, as shown, in the following tables:
•PLTPI (initialization) immediately after the DFHDELIM entry for CICS Transaction Server systems.
•PLTSD (shutdown) before the DFHDELIM entry.
CICS JCL changes
In this section, we provide the JCL changes that you can make.
Add a DD statement for RKANMOD and concatenate the Tivoli OMEGAMON II load library, RKANMOD, to DFHRPL:
//RKANMOD DD DISP=SHR,DSN=rhilev.RKANMOD
//DFHRPL DD DISP=SHR,DSN=......
// DD DISP=SHR,DSN=rhilev.RKANMOD
In addition to the changes above there are some optional JCL changes that you can make:
•//KOCGLBnn DD DUMMY
This instructs OMEGAMON CICS that the Global Data area with the suffix nn is to be used for this CICS region.
•//OCCIREQ DD DUMMY
This instructs OMEGAMON to check for the presence of the KOCCI address space. If it is not found, a message OC0806 is issued asking the operator if this CICS region is to continue in this case.
It is possible to run more than one copy of a KOCCI or OMEGAMON CICS agents on an LPAR. If this is the case, you must indicate which address space is to monitor this CICS region. This is achieved by placing corresponding DD cards in the CICS region JCL and the KOCCI address space JCL or agent JCL.
If more than one KOCCI is required, place the following card in both the CICS region and the KOCCI JCL:
//RKC2XMnn DD DUMMY
Where nn is a number between 00 and 15. If no card is specified, it is the same as using the number 00. Only one KOCCI can be active on an LPAR for a given release with any one number.
If more than one agent is required, put the following card in both the CICS region and the AGENT JCL:
//RKCPXMnn DD DUMMY
Where nn is a number between 00 and 15. If no card is specified it is the same as using the number 00. Only one agent can be active on an LPAR with any one number
Configuring OMEGAMON II for CICS
You configure OMEGAMON II for CICS using the Installation and Configuration Assistance Tool that is provided with ITM. Figure 10-30 on page 264 shows the configuration page for OMEGAMON II for CICS.
|
---------------- CONFIGURE OMEGAMON II FOR CICS / RTE: SC66 ------------------
OPTION ===>
Last selected
Perform these configuration steps in order: Date Time
I Configuration information (What's New)
1 Specify configuration values 08/09/18 13:35
2 Allocate additional runtime datasets 07/08/22 15:19
3 Create runtime members 08/08/28 16:55
4 Complete the configuration 08/09/11 22:08
Optional:
5 Allocate/initialize task history datasets 08/09/11 21:35
6 Manage CICS global data area(s) 08/09/16 15:08
7 Modify menu system command security 08/08/28 16:56
8 Install OMEGAMON Subsystem 07/08/22 15:20
F1=Help F3=Back
|
Figure 10-30 OMEGAMON II for CICS configuration panel
To configure OMEGAMON II for CICS:
1. Option 1, Specify configuration values: Walks you through the product-specific values for OMEGAMON II for CICS.
Figure 10-30 shows the first panel that is presented for option 1.
|
---------- OMEGAMON II FOR CICS CONFIGURATION VALUES / RTE: SC66 --------------
OPTION ===>
VTAM information:
Maximum number of CUA users ==> 99 (10-256)
Enable ACF/VTAM authorized path ==> N (Y, N)
CUA security options:
Specify security ==> RACF (RACF, ACF2, TSS,
NAM, None)
Function level security resource class ==> $OMEG (ACF2 max is 3 char)
Started task:
End-to-End (ETE) ==> started_task_name
Advanced options:
Maximum number of KOCCI users (UMAX) ==> 99 (1-99)
Copies of OMEGAMON II address spaces ==> 1 (1-16)
Fold CUA output to upper case ==> N (Y, N)
Enable CUA simplified signon ==> Y (Y, N)
Enable CUA WTO messages ==> N (Y, N)
Enter=Next F1=Help F3=Back
|
Figure 10-31 OMEGAMON II configuration values
a. Enter the name of the started task for the End-to-End component. Make sure that the name of the started task for End-to-End (ETE) is correct and is unique for this RTE. You can accept the defaults for the remainder of the values.
|
---------- OMEGAMON II FOR CICS CONFIGURATION VALUES / RTE: SC66 Row 1 from 48
COMMAND ===>
Modify the parameters below to suit your site's requirements.
ID: 00 Menu STC: CANSOC0_ Menu applid: &SYSNAME.OC0______________
CUA STC: CANSC20_ CUA applid: &SYSNAME.C20______________
CUA operator: &SYSNAME.C20O_____________
VTERM prefix: &SYSNAME.C0____________
Default CICS rgn: *_______
Major node: &SYSNAME.C20N_____________
Enter=Next F1=Help F3=Back F7=Up F8=Down
|
Figure 10-32 OMEGAMON II configuration values continued
In Figure 10-32, the started task names and applids are entered. In this case, we are using the z/OS system variables to prefix the applids, which makes it easier to configure environments for multiple LPARs.
2. Option 2, Allocate additional runtime data sets: You can ignore this step at this point because OMEGAMON II for CICS has no additional data sets allocated in this step.
3. Option 3, Create runtime members: Generates a JOB that can be submitted to create the required configuration members. This Job runs with a return code zero.
4. Option 4, Complete the configuration: Provides a checklist of the steps that you must complete. This describes such tasks as copying the started task JCL to the system proclib libraries, authorizing the required libraries, moving the VTAM definitions to the correct libraries, and installing the components in the CICS regions.
This completes the required steps to configure OMEGAMON II for CICS. The following optional steps are necessary, depending upon the options you want to use and the other products installed in the system.
1. Option 5, Allocate/initialize task history data sets: Is required if you require task history data to be maintained after a CICS region is shutdown.
2. Option 6, Manage CICS global data area(s): Where you can create new or edit existing global data areas. The global specifies the following:
– The features that are to be active for a CICS region.
– The group definitions for response time analysis and bottleneck analysis.
– The rules that are active for the resource limiting feature.
– The location of the task history data.
When you edit a global data area, there is a comprehensive help system that is available that describes each option in detail. This help is available by pressing PF1. The help is context sensitive and displays information based upon the location of the cursor. Upon exiting the edit, the global data area is validated and any errors are displayed.
After you make the changes the global data areas, copy them to the RKANPARU data sets. Selecting the option to copy global definitions generates JCL to move the members. This JCL specifies DISP=OLD for the RKANPARU data set. If you do not want to shut down all of the tasks that use this library, change this to DISP=OLD. The Job will not run until you exit the configuration tool completely to free up the data sets that are allocated to your TSO ID.
3. Option 7, Modify menu system command security: Use to specify the type of security that is to be active for the menu system commands. You can use it to set passwords for authorized commands, if an external security manager is not to be used. You can also use it to specify the level of security that is required for specific commands.
4. Option 8, Install OMEGAMON subsystem: Only required if this was not done for another OMEGAMON product on this LPAR.
OMEGAMON XE for CICS features
OMEGAMON XE for CICS provides a number of distinct features to allow you to monitor their CICS regions. Although much of the data that is provided shows the current state of resources or the usage of such resources. Some of the features use the collected data to provide real time or near real time summaries of the data that is collected, which provides users with extra insight into the performance of their CICS regions without having to run batch jobs to process large amounts of data.
Resource monitoring
The majority of the workspaces that OMEGAMON XE for CICS provides fall into the category of resource monitoring. They provide information that relates to the current state and usage of the resources that are available to the CICS region, which includes resources that are implicit, such as storage and those that are defined, such as files, programs, and so on.
Resource monitoring primarily allows for the creation of situations that allow the user to be notified when resources are unavailable, resource consumption exceeded certain thresholds, or available resources are below safe values. Careful tuning of the situations on a system can help to ensure that support staff are notified before the system availability is threatened.
OMEGAMON CICS currently provides information relating to the following resources in a CICS region:
•Automatic Initiate Descriptors (AIDs)
•Business Transaction Services process type details
•Business Transaction Services process details
•Business Transaction Services containers
•MRO connections
•ISC connections
•IP connections
•DB2 subsystem
•DB2 transactions
•DB2 thread activity
•Database control for IMS
•Dispatcher Summary
•Dispatcher TCB pools
•Dispatcher TCB modes
•Dump details
•Enqueue contention
•Enterprise Java Corba servers
•Enterprise Java request models
•Enterprise Java deployed java programs (DJARs)
•Exit programs
•File control data sets
•CICS region data sets
•Interval control elements
•Java programs
•CICS Journals
•CICS JVMs
•JVM profiles
•JVM classcaches
•Log streams
•LSR pools
•MQ status
•MVS TCBs
•Recovery manager Unit of work links
•Storage usage
•Storage DSA usage
•Storage Subpool usage
•System initialization values
•Transaction classes
•TCP/IP activity
•TCP/IP services
•Temporary storage
•Temporary storage queues
•Terminal storage violations
•Transactions
•Transient data
•Transient data queues
•Transaction manager
•UOW disposition
•UOW enqueues
•VSAM files
•VSAM RLS conflicts
•Web services
•Web services virtual hosts
•Web services pipelines
•Workrequests
In addition to the this list, there is the Region Overview query and workspace that provides information from several places in one query to allow for many important metrics to be returned in a single query. Region overview provides information, such as the transaction rate, the percent of maximum tasks, the i/O rate, CPU rate, and other details.
Service level analysis
Service level analysis is a great way to determine if transactions are meeting their performance objectives. It provides details relating to whether transactions are meeting their performance objectives. The objectives can be based upon the average response time or by percent of transactions that meet their goal response time.
Service level analysis produces summaries at the LPAR level. You can get the report selecting the CICS group node for an LPAR on the physical tree, as shown in Figure 10-33.
Figure 10-33 CICS Group node
The Service Level Summary report appears on the bottom pane of the workspace.
Figure 10-34 Service Level Summary report
The Service Level Summary report shows the summary for all of the transactions in all of the CICS regions for an LPAR that match the service Level analysis definition.
Selecting the link to Service Class Summary produces the following workspace in Figure 10-35.
Figure 10-35 Service Class Summary workspace
The Service Class Summary workspace shows two reports:
•To the right of the window is the Service Class by Region Summary, which provides a summary of all the CICS regions where any tasks that were classified in the service class run.
•The lower portion of the window is the Service Class by Transaction, which shows all the transaction IDs that run that were classified in this service class. In this case, there is only one, CPIH, but there is no restriction on the number of transactions that can be part of a service class.
The classification rules used by OMEGAMON XE for CICS to produce the summaries can either be those defined to z/OS Workload Manager or defined in OMEGAMON XE for CICS.
Configuring Service Level Analysis
To configure OMEGAMON XE for CICS you must first enable the CICS SLA view by adding the CICS SLA view to the assigned views for your user ID:
Figure 10-36 Administer Users icon
2. Select your current USERID and the Navigator Views tab, as shown in Figure 10-37.
3. Ensure that CICS SLA is in the Assigned View box. If it is not, in the available views box, click CICS SLA, and then click the left arrow.
Figure 10-37 Administer Users Navigator Views
4. After the CICS SLA view is added to the user, you can select it from the navigator pane in the TEP, as shown in Figure 10-38 on page 270.
Figure 10-38 Selecting CICS SLA view
Figure 10-39 shows the CICS SLA view.
Figure 10-39 CICS SLA view
To the right of the view we see the CICSplex Control information and the default Service Policy of DFLTSPOL.
In the CICSplex Control we configure how the classification is to take place:
•z/OS WLM: Indicates that the WLM assigned service class name will be used to classify the transactions.
•OMEGAMON WLM: Indicates that the rules that are defined using this interface determine the service class names that the transactions are classified in.
•Both: Indicates that the z/OS WLM classification will be used, if available. If the service class cannot be determined this way, the OMEGAMON rules classify the task.
The collection interval determines how often response data for service classes, CICS regions, and individual transactions is summarized and reported through the Service Class Analysis workspace.
A service policy applies to all service classes that are defined for your enterprise. Service-class goals vary by service policy: Service policies let you override a service class's response-time goals as dictated by your site's varying requirements, for example, you can have one service policy for prime-shift operation, another for nighttime operation, and a third for weekend operation.
To the left of the view we have service classes and workload groups. A workload group is one or more service classes that you to monitor as a unit. With it, you can monitor related service classes and highlight the worst-performing class within the group.
A service class identifies a block of related transactions that share common response-time goals for a single workload. The transactions must be related by transaction name, the user ID that invoked them, the VTAM terminal ID (LU name) that submitted them, the CICS region running them, or any combination thereof.
To define a new Service Class, Figure 10-40:
1. In the Service Class pane, click Create.
Figure 10-40 Define new Service Class window
2. Within the service-class editor, define a name for your new service class. The name you supply is converted to uppercase. Pull down the list of available workloads, and select the existing workload that you want to associate this service class with.
3. Within the Response Time pane, specify the response time that you expect for the CICS transactions that are associated with this service class.
4. Within the Goal pane, select one of these options:
– Activating the Average radio button means that the average response time for all matching transactions must at least meet the Response Time specified % of Goal
– Activating the % of Goal radio button and specifying a percentage mean that the given percentage of matching transactions must at least meet the Response Time specified
5. Press OK to accept the values.
6. Next you must create rules that control which transactions are classified as a part of this service class. Highlight the service class, and press Edit Rule. Figure 10-41 shows the rule definition window.
Figure 10-41 Rule definition window
7. Right-click a rule in the left pane, and a submenu is displayed, as shown in Figure 10-42.
Figure 10-42 Rule pop-up menu
8. Select either:
– Create TranID, to select based upon Transaction ID.
– Create UserID, to select based upon the user ID that is associated with the transaction.
– Create CICSname, to select based upon the JOB name of the CICS region.
– Create LUname, to select based upon the VTAM LU name for the terminal where the transaction originated.
A pop-up is then displayed where you can enter a value, as shown in Figure 10-43 on page 273.
Figure 10-43 pop-up window to enter rule value
9. Repeat these steps for other values, as shown in Figure 10-44.
Figure 10-44 completed rules
In this case, all transactions that are running in CICS regions, with job names starting with the characters A6P and having the transaction ID of CPIH, are classified in this service class.
Bottleneck Analysis
Bottleneck Analysis is a useful tool for people who want to improve the performance of the applications that run on a CICS region. Getting the most performance out of a CICS application can be compared to tuning almost anything, including a car! Extracting the maximum performance involves maximizing the time that is spent on productive work and minimizing that which puts a strain on the system.
In computer software terms, this means to minimize the time that an application must wait for something. If an application spends the bulk of its time waiting for a resource, such as storage, or a LSR buffers, then the application is not providing the maximum performance. In addition to elongated response times, the transactions hold resources themselves for longer and therefore potentially add to the contention occurring in the system.
Bottleneck Analysis helps you identify the wait reasons that your applications are experiencing. Every task in the system has a wait reason identified. That wait reason might be Running, in which case the task is doing productive work. Bottleneck Analysis has a list of wait reasons that it understands. Bottleneck Analysis runs as a subtask of the KOCCI. It periodically scans all of the tasks in the CICS region and accumulates counts for each of the wait reasons. Wait reasons that it knows nothing about are accumulated in an UNKNOWN bucket. If a task is in a wait reason that is turned off in the Bottleneck Analysis table, it is not counted.
By carefully selecting which wait reasons are active for a given system, it is possible to see those wait reasons that are truly impacting your applications and therefore what resources your applications are waiting for.
Figure 10-45 on page 274 shows an example of a Bottleneck Analysis.
Figure 10-45 Bottleneck Analysis example
In the example in Figure 10-45, there are three wait reasons that were identified:
•IC EXEC CICS DELAY: An application issue because it indicates that the task is explicitly waiting for a given period of time. There is not much from a performance stand point that you can do to improve this.
•XM MXT HELD: Indicates that 18 percent of the recorded wait reasons were for tasks that were delayed because of MAXTASKS. This might be because the value is too low.
•SM STORAGE REQUEST: Indicates that nearly 60 percent of the samples show that transactions are waiting for storage requests to be honored, which is the largest wait reason, so examine it first. It might be that the DSALIMIT is too low or that the application will simply not fit in one CICS region and some of the workload needs to be spread across other regions.
The same information can be displayed in the menu system, as seen in Figure 10-46 on page 275.
|
________________ ZBPDEX VTM A6POC3C1 V560./C SC66 09/22/08 16:15:53
> PF1 Help PF3 Back PF4 Main Menu PF7 Up PF8 Down PF11 Zoom
> A-Bottleneck Graph B-Bottlenecks C-Group Graph D-Group Bottlenecks
> E-Impact Analysis F-Impact Profile G-Impact Detail H-Enqueues
===============================================================================
> BOTTLENECKS
PDEX
+
+ Resource Resource
+ Type Name Short Term Information Long Term Information
+ -------- -------- ---------------------- ----------------------
+ % 0_______ 50_______100 % 0_______ 50_______100
+ CDSA (varies) 59 |------====> . .| 59 |------====> . .|
+ ICWAIT *TOTAL* 16 |--> . . . .| 16 |--> . . . .|
+ (none) (16) |--> . . . .| (16) |--> . . . .|
+ MXT XM_HELD 18 |--> . . . .| 18 |--> . . . .|
+
+ Samples . . : 43318 Samples . . : 43318
+ Elapsed . . : 5:59 MN Elapsed . . : 5:59 MN
+ Interval . : 10:00 MN Interval . : 30:00 MN
===============================================================================
|
Figure 10-46 Bottleneck Analysis in the Menu system
Controlling Bottleneck Analysis
Bottleneck Analysis is controlled through the global data area. The enablement of the feature is controlled in the <STARTUP_CONTROL> section, as shown in Example 10-1.
Example 10-1 Enabling Bottleneck Analysis
*
<STARTUP_CONTROL>
*
BOTTLENECK_ANALYSIS=AUTO
The BOTTLENECK_OPTIONS section specifies the parameters for the subtask:
*
<BOTTLENECK_OPTIONS>
*
CLEAR_INTERVAL_LONG=30
CLEAR_INTERVAL_SHORT=10
SAMPLE_INTERVAL=20
VARIABLE_BUCKETS=1000
EXCLUDED_TRANS=(CSSY,CSJC,CVST,CSSX,CSGX,CSNC,DSNC,CFQ*,KD4*)
Notes about the <STARTUP_CONTROL> section of Example 10-1:
•The CLEAR_INTERVAL_LONG and CLEAR_INTERVAL_SHORT controls the time span for the long and short term accumulations. Providing long and short term displays makes it possible to determine if a particular wait reason is a sudden change in the profile of an application or a longer term trend.
•The SAMPLE_INTERVAL value specifies, in tenths of a second, how often the subtask samples the active CICS region. Bottleneck Analysis can be a heavy user of CPU. The interval by default is set to two seconds. However in a busy system, this might be too high. Try to increase this value while still attempting to provide a statistically valid sample. It serves no purpose to make the interval such that only a few samples are accumulated for each wait reason.
•The VARIABLE_BUCKETS value specifies how many slots are reserved to Wait reasons that have a variable portion to their name. Bottleneck Analysis does not expand the slots that are available, so if a value of 1000 is specified, and more variable wait reasons than this are encountered, some information is lost. The default here is normally sufficient.
•The EXCLUDED_TRANS keyword specifies those transactions that are to be ignored by Bottleneck Analysis. It serves no purpose to collect information about system tasks and background server transaction that are permanently active, unless that is the workload you are interested in tuning.
•The last item of control is the list of wait reasons, Figure 10-47, which you can see by using the menu system option O.J.
|
________________ ZCDLST VTM CIWSS3C2 V560./C SC66 09/22/08 17:03:39
> PF1 Help PF3 Back PF4 Main Menu PF7 Up PF8 Down PF11 Zoom
> A-RTA On B-RTA Off C-RTA Status D-RTA Intervals E-RTA Scaling
> F-ONDV On G-ONDV Off H-ONDV Status I-Bottleneck Ctl J-Wait Reasons
> K-INTR Ctl L-IANL On M-IANL Off N-IANL Settings O-IANL Groups
> P-Collection Q-Shutdown R-RLIM On S-RLIM Off T-RLIM Status
> U-SMF Status V-ATF Filters W-ATF Status
===============================================================================
> CONTROL BOTTLENECK ANALYSIS WAIT REASON BUCKETS
BLST
+ BLST On/ Resource Resource Issuing Wait Reason Wait
+ ID Off Type Name Module Description Type
+ ---- --- -------- -------- -------- ------------------------ --------
: SY1W OFF (none) (none) DFHDUIO DU: Dump dataset I/O Systasks
: SY2W OFF (none) (none) DFHTISR TI: Timer service rq Systasks
: RMSL ON (none) (none) DFHRMSL7 RM: Keypoint process Systasks
: ZNAC ON (none) (none) DFHZNAC ZC: Terminal error Systasks
: DLCN ON (none) DLCNTRL DFHDBCT DBCTL: Work element DBCntl
|
Figure 10-47 Menu system wait reason control
Certain wait reasons are turned off for Bottleneck Analysis monitoring by default. If you determine that a wait reason is active and that you want to disable it, you can temporarily achieve this by changing its On/Off value or on a more permanent basis by coding the BOTTLENECK_ANALYSIS section in the global as shown in Example 10-2.
Example 10-2 Coding the BOTTLENECK_ANALYSIS section
*
<BOTTLENECK_ANALYSIS>
*
DSDF=NO
SODM=NO
SMRE=NO
Transaction History
It is often very useful to look back at recent transactions to see, in detail, the response time for a particular transaction or possibly look at resource consumption for an individual task. Although CICS SMF data provides the detailed information for a transaction, it can be a cumbersome task to gain access to recent SMF data and run an ad hock report.
OMEGAMON XE for CICS provides a feature called Transaction History, also known as Online Data Viewing (ONDV). This feature creates a data store for each CICS region and manages the space to hold as much transaction data as possible. Transaction History stores comprehensive information about a transaction with detailed file and database requests are collected.
You can select Transaction History by using the Online Data Viewing workspace under Transactions Analysis for a CICS region in OMEGAMON XE for CICS, as shown in Figure 10-48.
Figure 10-48 Transaction history display in TEP
The Menu system provides an equivalent display, as shown in Figure 10-49.
|
________________ ZONDV VTM A6POC3C1 V560./C SC66 09/22/08 18:40:02
> PF1 Help PF3 Back PF4 Main Menu PF7 Up PF8 Down PF11 Zoom
> A-History Record View B-History Record Selection
> C-Trace Record View D-Trace Filters Management
===============================================================================
> HISTORICAL TRANSACTION OVERVIEW
ONDV
+ Transaction Overview: 09/22/08
+ End Tran Task Term Type CPU Resp Storage File Term Abend
+ Time ID Num ID Time Time HWM Reqs I/O Code
+ -------- ---- ----- ---- ---- ----- -------- ------- ----- ---- -----
+ 16:03:09 CPIH 15713 n/a TRM .2 2.046357 33335K 3 0
+ 16:03:09 CPIH 15706 n/a TRM .2 1.980787 33335K 3 0
+ 16:03:09 CPIH 15711 n/a TRM .2 1.828366 33335K 3 0
+ 16:03:09 CPIH 15721 n/a TRM .2 1.544157 33335K 3 0
+ 16:03:09 CPIH 15720 n/a TRM .2 1.513379 33335K 3 0
+ 16:03:08 CPIH 15704 n/a TRM .2 1.558967 33335K 3 0
+ 16:03:08 CPIH 15699 n/a TRM .2 1.588833 33335K 3 0
+ 16:03:08 CPIH 15696 n/a TRM .2 2.098515 33335K 3 0
+ 16:03:08 CPIH 15695 n/a TRM .2 2.027101 33335K 3 0
+ 16:03:08 CPIH 15694 n/a TRM .2 2.195676 33335K 3 0
+ 16:03:08 CPIH 15679 n/a TRM .2 2.673689 33335K 3 0
+ 16:03:08 CPIH 15689 n/a TRM .2 2.423562 33335K 3 0
+ 16:03:08 CPIH 15690 n/a TRM .2 2.392998 33335K 3 0
+ 16:03:08 CPIH 15684 n/a TRM .2 2.577388 33335K 3 0
+ 16:03:08 CPIH 15686 n/a TRM .2 2.372017 33335K 3 0
|
Figure 10-49 Transaction History in the Menu system
The menu system also allows for the display of comprehensive detail information relating to a transaction. Selecting one of the tasks in the display in Figure 10-49 using the zoom key (PF11) provides the display in Figure 10-50 on page 278 and Figure 10-51 on page 279.
|
________________ ZZONDV VTM A6POC3C1 V560./C SC66 09/22/08 18:43:48
> PF1 Help PF3 Back PF4 Main Menu PF7 Up PF8 Down
===============================================================================
> HISTORICAL TRANSACTION DETAIL
ONDV 08
+
+ Task Detail Information
+
+ General Information
+ Transaction ID . . . . : CPIH Task number . . . . . . : 15696
+ Userid . . . . . . . . : CICSUSER Luname . . . . . . . . : None
+ Facility ID (local) . . : n/a Facility type (local) . : Term
+ Real transaction ID . . : CPIH Umbrella transaction ID : None
+ Program ID - first . . : DFHPIDSH Umbrella program . . . : None
+ Abend code . . . . . . : None
+
+ Time Statistics
+ CPU time . . . . . . . : 0.173600 Overall elapsed time . : 2.098515
+ Dispatch time . . . . . : 0.267648 Total wait time . . . . : 1.830848
+ Re-dispatch wait time . : 0.132016 Exception wait time . . : 0.000000
+ TS VSAM I/O wait time . : 0.000000 TD VSAM I/O wait time . : 0.000000
+ File I/O wait time . . : 0.000000 JC I/O wait time . . . : 0.000000
+ TC I/O wait time . . . : 0.000000 MRO wait time . . . . . : 0.000000
+ 1st dispatch delay time : 0.000112 Transaction class delay : 0.000000
+ Max tasks delay . . . . : 0.000000 Local ENQ delay . . . . : 0.000000
+ LU61 wait time . . . . : 0.000000 LU62 wait time . . . . : 0.000000
+ FEPI wait time . . . . : 0.000000 RMI elapsed time . . . : 0.000016
+ RMI suspend time . . . : 0.000000 RLS file I/O wait time : 0.000000
+ Syncpoint elapsed time : 0.000656 Lock manager delay . . : 1.288544
+ WAIT EXTERNAL wait time : 0.000000 WAITCICS and WAIT EVENT : 0.000000
+ Interval control wait . : 0.000000 Dispatchable wait time : 0.000000
+ Shared TS I/O wait time : 0.000000 RLS CPU time . . . . . : 0.000000
+ IMS wait time . . . . . : 0.000000 DB2 Readyq wait time . : 0.000000
+ DB2 Connection wait time: 0.000000 DB2 wait time . . . . . : 0.000000
+ SOCKET I/O wait time . : 0.505376 Global ENQ delay . . . : 0.000000
+ RRMS/MVS wait time . . : 0.000000 MAXOPENTCBS delay time : 0.000000
+ JVM elapsed time . . . : 0.000000 JVM suspend time . . . : 0.000000
+ QR TCB wait-for-dispatch: 0.012592 QR TCB elapsed time . . : 0.017008
+ QR TCB CPU time . . . . : 0.000720 Other TCBs elapsed time : 0.006112
+ Other TCBs CPU time . . : 0.005984 JVM(J8) TCB CPU time . : 0.000000
+ LE(L8) TCB CPU time . . : 0.166880 SS(S8) TCB CPU time . . : 0.000000
+ Program fetches wait . : 0.000000 Wait for a JVM TCB . . : 0.000000
+ JVM initialisation time : 0.000000 JVM reset time . . . . : 0.000000
+ Key 8 TCB elapsed time : 0.244512 Key 8 TCB CPU time . . : 0.166880
+ Key 9 TCB elapsed time : 0.000000 Key 9 TCB CPU time . . : 0.000000
+ J9 TCB CPU time . . . . : 0.000000 Wait for H8 TCB time . : 0.000000
+ RO TCB elapsed time . . : 0.000000 RO TCB CPU time . . . . : 0.000000
+ TCB mismatch time . . . : 0.000000 TCB change mode delay . : 0.036784
+ TCB create delay . . . : 0.000000 3270 Partner wait time : 0.000000
+
+ General Statistics
+ Primary term input msgs : 0 Primary term output msgs: 0
+ Primary term input chars: 0 Prmary term output chars: 0
+ Sec LU61 input msgs . . : 0 Sec LU61 output msgs . : 0
+ Sec LU61 input chars . : 0 Sec LU61 output chars . : 0
|
Figure 10-50 Menu system Task History detail
|
+ Sec LU62 input msgs . . : 0 Sec LU62 output msgs . : 0
+ Sec LU62 input chars . : 0 Sec LU62 output chars . : 0
+ TD gets . . . . . . . . : 0 TD puts . . . . . . . . : 0
+ TD purges . . . . . . . : 0 TS gets . . . . . . . . : 0
+ TS puts to aux . . . . : 0 TS puts to main . . . . : 0
+ Shared TS wait count . : 0 PC links . . . . . . . : 11
+ PC loads . . . . . . . : 0 PC xctls . . . . . . . : 0
+ JC writes . . . . . . . : 0 IC starts . . . . . . . : 0
+ Syncpoint requests . . : 1 BMS requests . . . . . : 0
+ BMS map requests . . . : 0 BMS in requests . . . . : 0
+ BMS out requests . . . : 0 CICS logger writes . . : 0
+ PC link URMs . . . . . : 0 IC requests . . . . . . : 0
+ 3270 bridge tran ID . . : n/a TS total requests . . . : 0
+ DPL requests . . . . . : 0 IMS/DBCTL requests . . : 0
+ DB2 requests . . . . . : 0 OO class requests . . . : 0
+ SSL bytes encrypted . . : 0 SSL bytes decrypted . . : 0
+ CICS TCBs attached . . : 0 TCB Mode Switches . . . : 110
+ WEB Receive requests . : 1 WEB Characters received : 0
+ WEB Send requests . . . : 0 WEB Characters sent . . : 0
+ WEB total request count : 12 WEB Repository READs . : 0
+ WEB Repository WRITEs . : 0 Local container bytes . : 0
+ Remote IC channel starts: 0 Remote IC channel data : 0
+ Total channel requests : 172 Browse channel requests : 0
+ Get container requests : 80 Put container requests : 88
+ Move container requests : 4 Total bytes for GETs . : 36710198
+ Total bytes for PUTs . : 31465325 DPL CHANNEL data bytes : 0
+ DPL RETURN CHNL bytes . : 0 LINK with CHANNEL . . . : 2
+ XCTL with CHANNEL . . . : 0 DPL with CHANNEL . . . : 0
+ RETURN with CHANNEL . . : 0 RETURN CHNL bytes . . . : 0
+ Client IP address . . . : 10.1.100.43
+ Tran Group ID (char) . : ..USIBMSC.A6POC3C1C.....zc..
+ Tran Group ID (hex) . . : 11EECCDEC4CFDDCFCFC0126AA800
+ 904292423B1676333139DD8F9300
+ TRGID: X'1910E4E2C9C2D4E2C34BC1F6D7D6C3F3C3F1C3091D2D68AFA9830000'
+
+ Storage Statistics
+ Getmains <16M . . . . . : 0 Getmains >16M . . . . . : 56
+ HWM <16M . . . . . . . : 0 HWM >16M . . . . . . . : 32553K
+ Occupancy <16M . . . . : 0K Occupancy >16M . . . . : 7570M
+ HWM of total pgm storage: 70K
+ HWM of pgm storage <16M : 0 HWM of pgm storage >16M : 70K
+ HWM pgm storage cdsa . : 0 HWM pgm storage ecdsa . : 0
+ HWM pgm storage rdsa . : 0 HWM pgm storage erdsa . : 40K
+ HWM pgm storage sdsa . : 0 HWM pgm storage esdsa . : 30K
+ Application Trace Active
+ File Control Statistics
+ Local Browses . . . . . : 0 Local Gets . . . . . . : 3
+ Local Adds . . . . . . : 0 Local Puts . . . . . . : 0
+ Local Deletes . . . . . : 0 Total Local Requests . : 3
+ Local VSAM calls . . . : 3 Total Remote Requests . : 0
+ Total Requests . . . . : 3
+
+ Umbrella Data
+ User work area (char) . : n/a
+ Unit-of-work Information
+ Netname . . . . . . . . : USIBMSC.A6POC3C1....
+ CICS token info (char) : ....n...
+ CICS token info (hex) . : 012B9100
+ 9DD35001
|
Figure 10-51 Menu system Task History detail continued
As you can see from Figure 10-50 on page 278 and Figure 10-51, a large amount of data is available for each task. Additionally, when the File Control Statistics heading is highlighted, it indicates that it is possible to zoom for more details.
Figure 10-52 on page 280 shows the file summary for a task.
|
________________ ZZONDVD VTM A6POC3C1 V560./C SC66 09/22/08 18:57:03
> PF1 Help PF3 Back PF4 Main Menu PF7 Up PF8 Down PF11 ZOOM
===============================================================================
> HISTORICAL FILE SUMMARY FOR SELECTED TASK
ONDV 08 FILE SUMMARY
+ Transaction Detail for CPIH task number =15696
+
+
+ File Control Statistics
+
+ Read . . . . . . . . . : 3 Read time . . . . . . . : 0.002944
+ Write . . . . . . . . . : 0 Write time . . . . . . : 0.000000
+ Update . . . . . . . . : 0 Update time . . . . . . : 0.000000
+ Delete . . . . . . . . : 0 Delete time . . . . . . : 0.000000
+ Browse . . . . . . . . : 0 Browse time . . . . . . : 0.000000
+ Misc request . . . . . : 0 Misc request time . . . : 0.000000
+ Total requests . . . . : 3
+
+ Database Requests Elapsed
+ Time
+ ---------------- -------- -------- --------
+ EXMPCONF 2 0.001792
+ EXMPCAT 1 0.001152
|
Figure 10-52 File summary for a task
Figure 10-53 shows that the transaction accessed two VSAM files. Again, you can get more detail on each file.
|
________________ ZZONDVD VTM A6POC3C1 V560./C SC66 09/22/08 19:01:03
> PF1 Help PF3 Back PF4 Main Menu PF7 Up PF8 Down
===============================================================================
> HISTORICAL FILE DETAIL FOR SELECTED TASK
ONDV 08 FILE DETAIL EXMPCONF
+ Transaction Detail for CPIH task number =15696
+
+
+ File Control Statistics
+
+ Database . . . . . . . : EXMPCONF
+ Read . . . . . . . . . : 2 Read time . . . . . . . : 0.001792
+ Write . . . . . . . . . : 0 Write time . . . . . . : 0.000000
+ Update . . . . . . . . : 0 Update time . . . . . . : 0.000000
+ Delete . . . . . . . . : 0 Delete time . . . . . . : 0.000000
+ Browse . . . . . . . . : 0 Browse time . . . . . . : 0.000000
+ Misc request . . . . . : 0 Misc request time . . . : 0.000000
|
Figure 10-53 File detail for a task
This level of detail shows, for each file accessed, the type of request and the response time to the application for those requests.
Controlling Transaction History
Transaction History is controlled through the global data area. The enablement of the feature is controlled in the <STARTUP_CONTROL> section, as shown in Example 10-3.
Example 10-3 Enabling the Transaction History feature
*
<STARTUP_CONTROL>
*
ONLINE_DATA_VIEWING=AUTO
The ONLINE_VIEWER section specifies the parameters for the subtask. Example 10-4 shows possible value configurations.
Example 10-4 Configuring values
*
<ONLINE_VIEWER>
DATA_STORE_TYPE=FILEOCMP
DATA_STORE_FILE_NAME=OMEGAXE.SC66.*.RKC2HIST
EXCLUDED_TRANS=(KD4O,KD4C,KD4,KD4D,CSSY)
*
or
*
<ONLINE_VIEWER>
DATA_STORE_TYPE=DSPACE
DATA_STORE_SIZE=956
RESERVED_SIZE=25
EXCLUDED_TRANS=(KD4O,KD4C,KD4,KD4D,CSSY)
Notes about Example 10-4:
•The DATA_STORE_TYPE keyword specifies how the data is to be stored. There are two options:
– FILEOCMP is where the data is stored to a VSAM Linear data set. This option allows the data to persist if the CICS region or the KOCCI is shutdown.
– DSPACE is where a data space is allocated to the KOCCI address space to hold the data. This option does not persist the values.
•The DATA_STORE_FILE_NAME is only applicable to a type of FILEOCMP. If the name provided contains an asterisk, as in Example 10-4, the asterisk is replaced with the JOB name of the CICS region.
•The DATA_STORE_SIZE and RESERVED_SIZE are only applicable to a type of DSPACE. The size refers to the size of the data space in kilobytes and the reserved size is the percentage of the space that is reserved for file details and application trace information.
•The EXCLUDED_TRANS keyword is applicable to both types and specified transactions that are not to be recorded to history.
Application trace
In addition to file and database detail statistics, OMEGAMON offers the capability to trace the application calls that a transaction makes. OMEGAMON traces calls made through the EXEC interface, the resource manager interface, and from third-party database providers, such as ADABAS, SUPRA, DATACOM, and IDMS.
The Application Trace display can be linked in OMEGAMON XE for CICS workspaces that show transaction History data. These are the Online Data Viewing, Units of Work, and the Web Services Transactions display. Figure 10-54 on page 282 shows the Application Trace workspace.
Figure 10-54 Application Trace workspace
The Application Trace workspace displays the trace data for a task to the right of the window. To the left, are details from the transaction history record.
Similar information is displayed in the menu system. You can display Figure 10-55 on page 283 by selecting the zoom key on the Application Trace Active heading in the transaction detail display. Zooming on the task from the Transaction History Trace record view also displays application trace (H.C).
|
________________ ZZONDA VTM CIWSS3C2 V560./C SC66 09/22/08 20:12:29
> PF1 Help PF3 Back PF4 Main Menu PF7 Up PF8 Down
==============================================================================
> TRACED TRANSACTION SUMMARY
ONDV 05 TRACE SUMMARY
+ Transaction Detail for ORDR task number =78877
+
+
+ Application Trace Facility
+ Lines 1 to 243 of 55
+ Trace Program Offset Function Resource Response Elapsed
+ Type Time
+ -------- -------- ------ ---------------- -------- ---------- --------
+ TSKSTRT DFHPIDSH 0 1ST DISPATCH
+ EXECIN CIWSMSGO 7DC GET CONTAINER 0.00006
+ EXECOUT CIWSMSGO 7DC GET CONTAINER NORMAL 0.00000
+ EXECIN CIWSMSGO 9CA GET CONTAINER 0.00000
+ EXECOUT CIWSMSGO 9CA GET CONTAINER NORMAL 0.00000
+ EXECIN CIWSMSGO A84 GET CONTAINER 0.00000
+ EXECOUT CIWSMSGO A84 GET CONTAINER NORMAL 0.00000
+ EXECIN CIWSMSGO B6E PUT CONTAINER 0.00000
+ EXECOUT CIWSMSGO B6E PUT CONTAINER NORMAL 0.00000
+ EXECIN CIWSMSGO FFFFFF GETMAIN 00004080 0.00000
+ EXECOUT CIWSMSGO FFFFFF GETMAIN 2824DC58 NORMAL 0.00000
+ EXECIN CIWSMSGO FFFFFF WRITEQ TD CESE 0.00000
+ EXECOUT CIWSMSGO FFFFFF WRITEQ TD CESE NORMAL 0.00000
+ EXECIN CIWSMSGO FFFFFF WRITEQ TD CESE 0.00000
+ EXECOUT CIWSMSGO FFFFFF WRITEQ TD CESE NORMAL 0.00000
+ EXECIN CIWSMSGO FFFFFF WRITEQ TD CESE 0.00000
+ EXECOUT CIWSMSGO FFFFFF WRITEQ TD CESE NORMAL 0.00000
+ EXECIN CIWSMSGO FFFFFF WRITEQ TD CESE 0.00000
+ EXECOUT CIWSMSGO FFFFFF WRITEQ TD CESE NORMAL 0.00000
+ EXECIN CIWSMSGO FFFFFF WRITEQ TD CESE 0.00000
+ EXECOUT CIWSMSGO FFFFFF WRITEQ TD CESE NORMAL 0.00000
+ EXECIN CIWSMSGO FFFFFF WRITEQ TD CESE 0.00000
+ EXECOUT CIWSMSGO FFFFFF WRITEQ TD CESE NORMAL 0.00000
+ EXECIN CIWSMSGO C52 GET CONTAINER 0.00000
+ EXECOUT CIWSMSGO C52 GET CONTAINER NORMAL 0.00000
|
Figure 10-55 Application Trace Menu System display
Application trace is a high overhead feature of OMEGAMON CICS. We recommend that you turn it on dynamically when required. Application trace can be controlled through the Control section of the Menu System. Specifically, you can use menu option O.W to activate and deactivate trace. Use menu option O.V to specify trace filters, which allow for trace to be active only for certain transactions.
Enterprise Java Monitoring
OMEGAMON XE for CICS provides monitoring for CICS Enterprise Java as part of the resource monitoring that we described in “Resource monitoring” on page 266. The primary workspace for monitoring is the Enterprise Java Analysis branch under a CICS region. Figure 10-56 on page 284 shows an example of the Enterprise Java Analysis workspace.
Figure 10-56 Enterprise Java Analysis workspace
From the Java Analysis workspace, you can see summary information that relates to the Enterprise Java resources in a particular CICS region. This workspace shows the following resources defined in CICS:
•Corbaserver Summary
•Request Model Summary
•DJAR Analysis
•Enterprise Java Bean Analysis
The Corbaserver and Request Model resources contain links to more details:
•For Corbaservers, the link displays more details about the resource, for example, the TCP/IP Service name and JNDI details.
•For Request Models, the link shows more details about the resource, such as Corbaserver, Transaction Identifier, Bean name, Interface, Module, and Operation.
Java Program Analysis
OMEGAMON XE for CICS provides monitoring for CICS Java Programs as part of resource monitoring. The primary workspace for monitoring is the Java Program Analysis branch under a CICS region. Figure 10-57 on page 285 is an example of the Java Program workspace.
Figure 10-57 Java Program Analysis workspace
From the Java Program Analysis workspace, you can see information that relates to the Java Program resources in a particular CICS region.
Of particular interest is the JVM class that is associated with the program.
JVM Analysis
OMEGAMON XE for CICS provides monitoring for CICS JVMs. The primary workspace for monitoring is the JVM Analysis branch under a CICS region. Figure 10-58 on page 286 shows an example of the JVM Analysis workspace.
Figure 10-58 JVM Analysis workspace
From the JVM Analysis workspace, you can see information that relates to the JVM resources in a particular CICS region. This workspace shows the following resources:
•Java Virtual Machine Analysis
•JVM Profile Analysis
•JVM Classcache Details
•JVMPool Statistics
The Java Virtual Machine Analysis pane has a Dynamic Workspace Link to the Single Transaction Analysis Workspace, shown in Figure 10-59 on page 287. The link is greyed out when a JVM is not associated with a transaction.
Figure 10-59 Single Transaction Analysis Workspace
This workspace is identical to the Transaction Analysis workspace. The only difference is that it is filtered using the task number of the transaction that is executing the JVM. All the Links that are associated with a transaction in the Transaction Analysis workspace are available to the single transaction instance.
Resource limiting
Resource limiting (RLIM) is a special feature that OMEGAMON CICS provides. This feature is designed to automatically protect the environment from rouge CICS transactions that might loop in a way that cannot be detected by the CICS runaway task protection or transactions that use extraordinary amounts of certain resources.
Resource limiting examines the values for certain resources or the number of certain types of requests. If the specified thresholds were exceeded, OMEGAMON either issues a message or causes the transaction to abnormally terminate (ABEND).
The enablement of the feature is controlled in the <STARTUP_CONTROL> section, as shown in Example 10-5.
Example 10-5 Enabling Resource limiting
*
<STARTUP_CONTROL>
*
RESOURCE_LIMITING=AUTO
RESOURCE_LIMITING_MSG_DEST=TDQ
RESOURCE_LIMITING_SYSTEM_TASKS=NO
RESOURCE_LIMITING_ABEND_CANCEL=YES
Notes about the options in Example 10-5:
•The option RESOURCE_LIMITING_MSG_DEST specifies if the message is to be written to the transient data queue CSSL (TDQ) or to the system console (LOG).
•RESOURCE_LIMITING_SYSTEM_TASKS allows you to specify if resource limiting is to issue warning messages for CICS system transactions. OMEGAMON will not cause a system task to ABEND.
•RESOURCE_LIMITING_ABEND_CANCEL allows you to specify whether any application abend handling exits are to be honored when RLIM elects to ABEND a transaction. Specifying NO indicates that ABEND exits will not be cancelled and therefore will remain in effect.
•The <RESOURCE_LIMITING> section of the global specifies the limits that are to be in effect for a given transaction ID, as shown in Example 10-6.
Example 10-6 <RESOURCE_LIMITING>
*
<RESOURCE_LIMITING>
*
<<CPU>>
INCLUDED_TRANS=(TRLN,DE*)
KILL_LIMIT=10
WARN_LIMIT=5
*
<<VSAM>>
INCLUDED_TRANS=(STRS,F?L?,TRLN)
KILL_LIMIT=100
WARN_LIMIT=50
*
In Example 10-6, transactions TRLN has the message OC8902 issued, if it makes more than 50 EXEC CICS requests against a FILE or if the CPU used exceeds five seconds. If the transaction exceeds 10 seconds of CPU or 100, EXEC CICS FILE requests the message OC8903 get issued and the transaction is abended.
Specifying a question mark (?) in the transaction ID indicates that the rule matches for a transaction ID that has any character in that location. So a transaction ID of FALZ has WARN and KILL limits of 50 and 100 for VSAM. Transaction ID FLAZ will not.
An asterisk (*) can be specified at the end of a transaction ID to indicate any trailing characters will match.
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.