《1229+将复制数据队列管理器迁移到原生高可用性 CRR 即插即用架构.pdf.pdf》由会员分享,可在线阅读,更多相关《1229+将复制数据队列管理器迁移到原生高可用性 CRR 即插即用架构.pdf.pdf(25页珍藏版)》请在三个皮匠报告上搜索。
1、1229 Migrating RDQM to Native HA CRRA Drop in ArchitectureReliability is not optional,its essentialRDQM as a Starting Point HA in cluster 2 Regions for DR Live primary Recovery-near or far data center/cloud region Single interface for application via“global IP”How to achieve this?RDQM DependenciesRe
2、quirementsHost Based Deployment on RHEL7+EC2 in AWSBare MetalVMAt least 3 nodes for quorumAdvanced MQ licensing for Failover and HA Replica licenses for extra nodesRDQM ImpedimentsGotchasDRBD and OS Kernel PatchingDRBD performance impact during a-syncNo“self healing”Limited to EC2/VM/Bare Metal moni
3、toring of the healthIBM MQ doesnt fully control the architecture as it relies upon RHEL for kernel elementsRDQM or Native HA CRR?Native HA CRR as a Replacement/Strategy HA in cluster 2 Regions for DR Live primary Recovery-near or far data center/cloud region Single interface for application via“glob
4、al IP”Same as beforeNative HA CRR DependenciesRequirementsKubernetes(pick your flavor)EC2 in AWSEKSOpenshiftBare MetalVMAt least 3 nodes for quorum(RAFT arch)Single Advanced MQ licenseNative HA CRR AdvantagesWhat works better?No longer dependent upon DRBD for synchronization and heart beat monitorin
5、gDRBD performance impact goes away for a-syncHA functionality based on RAFT modelKubernetes determines health of pods and failovers within cluster for HALive Demo?Native HA CRR Setup ExampleScreen Recording 2025-08-08 at 11.21.43AM.movNative HA CRR Performance&MetricsScreen Recording 2025-08-08 at 1
6、1.37.55AM.movNative HA CRR Write Persistent MessagesScreen Recording 2025-08-08 at 12.01.58PM.movNative HA CRR Failover Example/EvidenceScreen Recording 2025-08-11 at 7.54.01AM.movOf the runs on Native HA CRR,the following statistics are ta