System Programmer's Guide to: Workload Manager

1 CoverSystem programmer s Guide to: Workload ManagerPierre CassierAnnamaria DefendiDagmar FischerJohn HutchinsonAlain ManevilleGianfranco MembriniCaleb OngAndrew RowleyWorkload Manager overview and functionalitiesHow to classify your workloads Best practices samplesInternational Technical Support OrganizationSystem programmer s Guide to: Workload ManagerMarch 2008SG24-6472-03 Copyright International Business Machines Corporation 2005, 2006, 2007. All rights to Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP ScheduleContract with IBM Edition (March 2008)This edition applies to Version 1, Release 8 of z/OS (product number 5694-A01).

2 Note: Before using this information and the product it supports, read the information in Notices on page xi. Copyright IBM Corp. 2005, 2006, 2007. All rights .. xiTrademarks .. xiiPreface .. xiiiThe team that wrote this redbook.. xiiiBecome a published author .. xivComments welcome..xvChapter 1. Introduction.. Introduction to Workload management .. Basic concepts .. Service Level Agreement .. Workload Manager components .. Service definition and service policies .. Workload classification .. Service classes .. Importance of a goal .. Resource Group .. Application Environment.

3 Scheduling environment .. Sampling dispatchable unit states .. Performance management metrics .. 15 Average transaction response time .. 15 External throughput rate (ETR) .. 17 Resource utilization .. 18 Saturation design point (SDP) .. Processor performance metrics .. 20 CPU time .. 20 Cycle time .. 21 Millions of instructions per second (MIPS).. 21 MFLOPS .. 21 CPU Service Units: System Resource Manager (SRM) constant .. 22 Internal throughput rate (ITR) .. 22 Large Systems Performance Reference (LSPR) .. I/O performance and metrics.. 25 Chapter 2. How WLM works.. Workload management.

4 Z/OS dispatchable units (DUs) .. 30 WLM Business Unit of Work .. 31 The idea of dispatchable units .. 32 Enclave .. Understanding the algorithms .. 35 Workload management .. Performance index .. 36 Sysplex PI and local PI .. 36 Performance index evaluation in a sysplex .. Policy adjustment function .. 40 Impact of the number of service class periods.. 42iv System programmer s Guide to: Workload ManagerDonor and receiver determination .. 43 Selecting a receiver candidate.. 45 Find a receiver bottleneck .. CPU management .. 46 Dispatching priorities assignment .. 47 Discretionary goals .. Swap control.

5 51 Swap protect time .. Storage targets .. I/O management .. 52I/O queues .. 52 DASD I/O priority .. 53 TAPE I/O priority .. 53 Sysplex I/O priority management.. 53I/O priority management decisions .. 54 Channel subsystem I/O priority queuing (CSS IOPQ) .. 54 Dynamic Channel Path Management (DCM).. 55 Parallel Access Volumes (PAV).. Transactions and management.. How WLM treats work in the System .. Enclave management .. 64 Classification .. 64 Application Environment enclave server .. 64 Application Environment enclave server topology .. Execution delay monitoring .. 65 Monitoring environment or performance block (PB).

6 66 Services .. 66 Server topology .. 67 Server management .. 71 Transaction phase .. 71 Summary .. 73 Chapter 3. WLM functions .. Protecting work .. Resource Groups .. 76 Types of Resource Groups .. 76 Defining Resource Groups.. 77 Resource Group maximum management .. 78 Service ramps up .. 80 RMF reporting .. 81 Resource Group usage examples .. CPU critical .. 85 CPU CRITICAL .. 86 CPU CRITICAL in action .. 86 Who should use CPU CRITICAL .. 87 Where to specify CPU CRITICAL .. Storage critical .. 88 Who should use STORAGE CRITICAL .. 88 Where to specify STORAGE CRITICAL .. Initiator dispatch priority.

7 90 Initiator behavior .. WLM CPU preemption .. ENQ management .. Server address space management .. Application Environment .. 92 Contents vApplication Environment queues .. 93 Transaction flow in an Application Environment .. 93 Application Environment and transaction flow example.. Workload routing support .. VTAM generic resources.. 95 VTAM exit customization .. Sysplex routing .. 97 Calculation of server weight.. 98 Sysplex routing user: z/OS Load Balancing Advisor .. 99 Sysplex routing user: DB2 Connect .. Soft capping .. Defined capacity .. Group capacity .. Intelligent Resource Director.

8 102 Policy considerations for IRD.. 103 Importance and goal of a service class .. 103 Minimum number of logical processors .. Linux for System z LPAR management .. 103 Setting the cluster name .. 103 Setting the System name .. 103 Setting the goal .. Specialty processors .. Internal Coupling Facility Processor (ICF) .. Integrated Facility for Linux Processor (IFL) .. zSeries Application Assist Processor (zAAP) .. 106zSeries Application Assist Processor and Java code flow.. zSeries Application Assist Processor positioning .. WLM zAAP DUs states .. z9 Integrated Information Processor (zIIP).

9 109z9 Integrated Information Processor and DB2 code flow.. z9 Integrated Information Processor positioning.. WLM zIIP DU states .. 110 Chapter 4. Implementation and Workload classifications.. Overview .. WLM implementation .. IEAOPTxx usage .. WLM ISPF application .. 119 Security access .. 119 Starting your WLM ISPF application .. WLM couple data set management .. 121 Allocating the WLM couple data set.. 121 Managing your WLM CDS .. Specifying your service definition .. Service policy definition .. Workload specification .. Resource Group specification .. Working with service classes.

10 Classification rules .. Service definition coefficient .. Report class .. Application Environment .. Scheduling environment .. Policy overrides.. 142vi System programmer s Guide to: Workload How to use policy overrides .. 143 Adding and removing a Resource Group.. 143 Overriding a Resource Group definition.. Evaluating your SMF records .. Installing and activating your service definition .. Current WLM users upgrading to a new OS .. WLM compatibility and coexistence .. WLM enhancements matrix .. WLM tools .. Service Definition Formatter .. WLM Work Queue Viewer.