3147 - 可用性和性能：异步系统管理的双工模式和 Db2 for z_OS LOCK1 结构.pdf

1、Sponsored byData Technical Summit IBM TechXchange 2025Availability AND Performance:async system managed duplexing and Db2s LOCK1 structureSession 3147Async duplexing of Db2 for z/OS LOCK1 structureMark RaderIBM Z WSC Db2 for z/OS SpecialistDoes Parallel Sysplex and data sharing solve,or prevent,avai

2、lability problems?IBM TechXchange|2025 IBM Corporation2If we implement Parallel Sysplex and Db2 for z/OS data sharing(*)will we achieve 100%availability?Sysplex and data sharing provide the infrastructure upon which to build high availability solutions.One part of that infrastructure,the Db2 lock st

3、ructure,is the focus of this session.Agenda010203040506Parallel Sysplex conceptsStructure rebuild and duplexingSystem managed duplexing(SMD)Synch SMD vs.Async SMDUse cases for Async SMDCustomer experiencesIBM TechXchange|2025 IBM CorporationParallel Sysplex basics for Db2 for z/OS Multiple Db2 membe

4、rs comprise a Db2 data sharing group Coupling facility(CF)structures provide high-speed sharing of locks,status,and data(tables,indexes)Managed by cross-system extended services,or XES CFs run coupling facility control code(CFCC)ssnmIRLM allocates the lock structure:dsngrp_LOCK1 ssnmMSTR allocates t

5、he shared communication area:dsngrp_SCA ssnmDBM1 allocates a group buffer pool(GBP)for each local BP with shared data:dsngrp_GBPn GBPs not pictured;spread across both CF LPARsIBM TechXchange|2025 IBM Corporation4Parallel Sysplex basics for Db2 for z/OS LOCK1 and SCA are required structuresIf Db2 can

6、not allocate them,Db2 failsIf LOCK1 lost,all IRLMs required to rebuild LOCK1If SCA lost,all MSTRs required to rebuild SCARebuild to the alternate CF LPARIBM TechXchange|2025 IBM Corporation5Original Parallel Sysplex configuration;at least 4 systems 2 central electronic complexes(CECs)with production