版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领
文档简介
1、MySQL 8数据库复制技术MySQL 8.0: Whats New In ReplicationIntroductionUse CasesEnhancements in MySQL 8ConclusionProgram Agenda12342Introduction13Database ReplicationServerBServerAAppINSERT .INSERT .INSERT .Replication4“The process of generating and reproducingmultiple copies of data at one or more sites.”,Da
2、tabase Systems: A Practical Approach to Design, Implementation, and Management, Thomas M. Connolly, Carolyn E. Begg, Third Edition, 2002.MySQL Database Replication: OverviewINSERT .ServerBbinary logINSERT .relay logINSERT .ServerAbinary logAppReceiver Meta-data UpdateApplier Meta-data UpdateINSERT .
3、Comm.Framework5MySQL Database Replication: OverviewINSERT .ServerBbinary logINSERT .relay logINSERT .ServerAbinary logAppReceiver Meta-data UpdateApplier Meta-data UpdateINSERT .Persistent log bufferThreaded applierSend, Receive,ACK, NACK,Heartbeating,.Comm.FrameworkCapture statements or data change
4、s.Persistent log buffer6MySQL Database Replication: Some NotesBinary LogLogical replication log recording master changes (binary log).Row or statement based format (may be intermixed).Each transaction is split into groups of events.Control events: Rotate, Format Description, Gtid, and more.Layout of
5、 the Binary Log.GTIDBEGINE1E2.COMMITGTIDBEGINE1E2.COMMIT7MySQL Database Replication: Some NotesCoordination Between ServersABSince 3.23ABsemi-synchronous (plugin)ACBgroup replication (plugin)transactionstransactionsSince 5.5asynchronous (native)transactions, membership, coordination8acksSince 5.7.17
6、And now in MySQL 8 as of 8.0.1Use cases29Clustering Made PracticalReplicate Automate I ntegrate ScaleEnhance10ReplicateGroup ReplicationFor highly available infrastructures where:the number of servers has to grow or shrink dynamically;with as little pain as possible.ABCDE11AutomateGroup ReplicationS
7、ingle-primary modeAutomatic PRIMARY/SECONDARY role assignmentAutomatic new PRIMARY election on PRIMARY failuresAutomatic setup of read/write modes on PRIMARY and SECONDARIESAutomatic global consistent view of which server is the PRIMARYSSS*SPSNext PrimaryOld Primary12IntegrateBinary LogLogical repli
8、cation logExtract, transform and load.MySQL fits nicely with othertechnologies.INSERT .ServerAReplication logAppINSERT .HadoopKafkaSolrLucene.Transform13ScaleAsynchronous ReplicationReplicate between clustersFor disaster recoveryRead ReplicasFor read-scale out. Deploy asynchronous read replicas conn
9、ected to the clusterABCJKLDIZM.14InnoDB ClusterEnhanceS115S2S3S4SMMMApplicationMySQL ConnectorMySQL RouterApplicationMySQL ConnectorMySQL RouterMySQL ShellHAReplicaSet 1InnoDB Cluster Integrated SolutionGroup Replication for high availability.Asynchronous Replication for Read Scale-out.One-stop shel
10、l to deploy and manage thecluster.Seamlessly and automatically route the workload to the proper database server in the cluster.Hide failures from the application.Enhancementsin MySQL 3.4Binary Log Metadata Operations Monitoring PerformanceOther3.516New Metadata in the Binary LogEasy to ex
11、tract, transform and load into other systems.ServerAGtid_log_event Transaction length, Commit timestamps Table_map event Types, primary key, Sign, charset, Write_rowsCommitBinary logAppINSERT .17New MetadataEasy to decode what is in thebinary log.Reduces impedance while connecting MySQL to other sys
12、tems using the binary log.Capturing data changes throughthe binary log is simplified.Also more stats showing where the data is/was at a certain point in time.Enhancementsin MySQL 3.4Binary Log Metadata Operations Monitoring PerformanceOther3.518Multi-Source Replication FiltersReplicate, F
13、ilter, Aggregate, QueryABCusersbiogroupsuserspostscommentsusersbiogroupspostscommentsonly users19Preventing Updates On Replicas that Leave the ClusterAutomatic protection against involuntarily tainting of offline replicasBCABCAA joins a single primary cluster and it allows update operations prior to
14、 joining.A becomes a secondary and is automatically set to read only.BCAA leaves the group and remains read only until the DBA reverts.20Primary Election WeightsChoose next primary by assigning election weights to the candidates.Bw=100Cw=90Aw=50Bw=100Cw=90Aw=50Cw=90Aw=50B is the primary.B is not in
15、the group anymore.C has higher weight than A. C is elected the new primary.21AppsAppsOnline and Automatic Cluster Performance ManagementNew options to fine tune the cluster automatic flow control in 8.0.BCABCABCAB is struggling to keep up.BA iasnsdtrCugthglriontgtlteo themselvkeeseapuutopm. atically
16、.B is not struggling any more.System adjusted Itself.B has a chance to catch up.Apps22Enhancementsin MySQL 3.4Binary Log Metadata Operations Monitoring PerformanceOther3.623Monitor Lag With Microsecond PrecisionThrough the entire asynchronous topologyABCDHow much time does my data take to
17、 reachD coming from A?24Monitor Lag With Microsecond PrecisionFrom the immediate masterABCDHow much time does my data originated inA takes to flow from B to C?25Monitor Lag with Microsecond PrecisionFor each stage of the replication applier processDDataWorker ThreadsRelay LogCoordinator ThreadReceiv
18、erThread26Per Stage Timestamps User can monitor how much time it takes for a specific transaction to traverse the pipeline.Version, Role and moreGlobal Cluster Stats Available on Every ServerABCA is primary.B, C are secondaries. A: 5.7.20B, C: 8.0.3A is primary.B, C are secondaries. A: 5.7.20B, C: 8
19、.0.3A is primary.B, C are secondaries.A: 5.7.20B, C: 8.0.327Query one Replica, Get statusof allEvery replica reports cluster-wide information about roles and versions of the members of the cluster.Also available at any replica arecluster-wide status.Enhancementsin MySQL 3.4Binary Log Meta
20、data Operations Monitoring PerformanceOther3.528Highly Efficient Replication ApplierWrite set parallelization5000045000400003500030000250002000015000100005000012464128256Updates/second Applies on the Replica81632Number of Clients on the MasterApplier Throughput: Sysbench Update IndexCOMMIT_ORDERWRIT
21、ESETWRITESET_SESSION29Highly Efficient Replication Applier30Write set parallelizationWRITESET dependency tracking allows applying a single threadedworkload in parallel. Delivers the best throughput of the three dependency trackers, at any concurrencylevel.WRITESET_SESSION in addition to writesets tr
22、acks sessions dependencies as well. Two transactions executed on the same session are always scheduled in execution order on replica servers.Fast Group Replication recovery time to catch up.Replica quickly online by using WRITESETFast Group Replication Recovery109876543210Sysbench RW at 33% capacity
23、 (workload: 9K TPS on 64 threads)Sysbench RW at 66% capacity (workload: 18K TPS on 64 threads)Time to cacth-up per time of worload missing (ratio)Group Replication Recovery Time: SysbenchUpdate Index (durable settings)MySQL 5.7.20MySQL 8.0.3109876543210Sysbench RW at 33% capacity (workload: 4K TPS o
24、n 64 threads)Sysbench RW at 66% capacity (workload: 8K TPS on 64 threads)Time to cacth-up per time of worload missing (ratio)Group Replication Recovery Time: SysbenchRW (durable settings)MySQL 5.7.20MySQL 8.0.331High Cluster ThroughputMore transactions per second while sustaining zero lag on any rep
25、lica8641632Number of Clients on the MasterAsynchronous Replication Sustained Throughput(Sysbench Update Index, non-durable settings)MySQL 5.7MySQL 8.0.3864Updates per second1632Number of Clients on the MasterAsynchronous Replication Sustained Throughput(Sysbench Update Index, durable settings)MySQL
26、5.7MySQL 8.0.345000400003500030000250002000015000100005000032High Cluster Throughput33More transactions per second while sustaining zero lag on any replicaAt lower thread count, the throughput of the system doubles in MySQL 8.0 comparedto MySQL 5.7 on durable settings.At lower thread count, the thro
27、ughput of the system more than doubles in MySQL 8.0compared to MySQL 5.7 on non-durable settings.Replicate only changed fields of documents (Partial JSON Updates)Efficient Replication of JSON Documents2000010000030000400005000060000Entire JSONPartial JSONBytes per transactionBinary Log Space Per Tra
28、nsactionFULLMINIMAL34Numbers are from a speciallydesigned benchmark:tables have 10 JSON fields,each transaction modifiesaround 10% of the data0.0500.01000.01500.02000.02500.0Transactions per secondThroughput on the Master: Partial JSON vs Complete JSON48163264Replicate only fields of the document th
29、at changed (Partial JSON Updates)Efficient Replication of JSON Documents0100020003000500040006000700080009000Throughput on the Slave: Partial JSON vs Ccomplete JSON48163264F1ULLMIN2IMAL COMPLETE JSONF3ULLMIN4IMAL PARTIAL JSONFULLMINIMAL COMPLETE JSONFULLMINIMAL PARTIAL JSON35Enhancementsin MySQL 83.
30、Binary Log Metadata Operations Monitoring PerformanceOther3.536Changes to defaults in MySQL 837High performance replication enabled out-of-the-boxBinary log is on by default.Logging of slave updates is on by default.Replication metadata is stored in InnoDB tables by default instead of fil
31、es.Row-based applier uses hash scans to find rows instead of table scans.Transaction write-set extraction is on by default.Binary log expiration is set to 30 days by default.Server-id is set to 1 by default instead of 0.Other MySQL 8 Replication Enhancements38Monitoring: Monitor replication even whe
32、n disk fullMonitoring: Current query being applied, even for row-based replicationMonitoring: Replication filters statistics in performance schemaMonitoring: Group Replication threads instrumented and shown inperformance schemaRecoverability: Recover DDL and binary log together after a crashOther My
33、SQL 8 Replication Enhancements39Operations: Restore global transaction identifiers metadata on a non-empty serverOperations: Specify binary log file number after RESET MASTEROperations: Specify when binary log files are automatically purged (withsecond precision)Operations: SAVEPOINT support when wr
34、ite sets are being extractedTroubleshooting: Dynamic and high performance debugging of groupreplication inter-node messagingRoadmap340GR 0.8.0 labs (beta)replication stream compre sionSSL supportIP whitelistingerror log enhancementssplit brain handlingThe Road to MySQL 8 Group Replication and InnoDB ClustersMySQL 5.6.10GR 0.2.0 labsHello world!GR 0.3.0 labssupport for corosync 2.xServer side changesGR 0.4.0 labsversion handling.cauto-infields handli grecovery enha
温馨提示
- 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
- 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
- 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
- 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
- 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
- 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
- 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。
最新文档
- 2026年国际阴郁症心理测试题及答案
- 2022上海事业单位统考历年真题+刷题组答案解析
- 2023年广西事业单位考试B类模拟题及答案 下载量超10万的备考资料
- 2026社招德语游戏客服3年经验面经配套面试题库及标准答案
- 2021临床器械试验方案设计专项考试题及详细答案解析
- 2024工地铆工安全考核必刷题及标准解析答案
- 2024中储粮笔试历年高频考题及标准答案解析
- 开美发店股东协议书
- 首发精神分裂症的治疗
- 整体护理病例健康指导
- 2026中国商用飞机公司招聘面试题库
- 4.1《致敬劳动者》课件 统编版道德与法治三年级下册
- 中考总复习数学100道基础题三大专题
- OpenClaw专题学习培训
- 安徽省合肥市一六八中学2026届高三3月份规范训练 语文试卷(含答案详解)
- 第一章 三角形的证明及其应用 单元测试(含答案)2025-2026学年数学北师大版八年级下册
- 2026年迎接国家义务教育质量监测工作实施细则方案及应急预案
- (2025年)食品生产许可证审查员考试全考点试题带答案
- 水包砂施工技术交底
- 国别与区域研究毕业论文
- 防水公司挂靠协议书
评论
0/150
提交评论