Scaling with PostgreSQL 9.6 and Postgres-XLTransaction Handling Storage XL Coordinator Parser Planner Executor “Metadata” Storage XL Datanode Executor Storage Postgres Conference China 2016 中国用户大会 Postgres-XL Technology • Users connect Proc 2 GTM Postgres Conference China 2016 中国用户大会 GTM Proxy • Runs alongside Coordinator or Datanode • Backends use it instead of GTM • Groups requests together • Obtain range of transaction ids (XIDs) read mainly tables • Sometimes good for dimension tables in data warehousing • If coordinator and datanode are colocated, local read occurs • Bad for write-heavy tables • Each row is replicated to all0 码力 | 87 页 | 1.16 MB | 1 年前3
2. 腾讯 clickhouse实践 _2019丁晓坤&熊峰… … app-n RPC DataNode NameNode MetaStore Controller manager Scheduler RPC app-1 Data1 Partition0 Data2 Partition2 DataN PartitionM … … app-2 … … app-n RPC DataNode 基于位图的分布式计算引擎 API API Server Scheduler SQL-Parser QueryOptimier Column1 DataNode Column2 Column3 ColumnN Column1 DataNode Column2 Column3 ColumnN bitmap 画像下钻分布式计算引擎 多维 提取 iData大数据分析引擎 分布式多维计算引擎 基于位图索引和行式内容存储 Execution Plan Bitcode Emitter Scheduler Block 1 Block 2 Block … DataNode-2 Block 1 Block 2 Block … DataNode-3 DataNode-1 指标计算平台 Ø 分布式计算 • 并行计算 Ø 列式存储 • 按需加载减少IO • 可支持大量列 Ø 动态位图索引 •0 码力 | 26 页 | 3.58 MB | 1 年前3
HBase最佳实践及优化HBase 集群Master和Zookeeper 节点 IBM 3650 PC,双路六核,Intel X5650处理器, 2.66GHz主频,48GB内存,6*1TB SATA硬盘 5台 DataNode/TaskTracker/Region Server IBM 3650 PC,双路六核,Intel X5650处理器, 2.66GHz主频,48GB内存,6*1TB SATA硬盘 69台 用户访问查询页面的时间) – 支持并发查询数目:1000请求/秒 Postgres Conference China 2016 中国用户大会 系统部署 • NameNode节点:3台 • DataNode(数据存储节点): 178台 • Zookeeper节点:7台 • 集群监控节点:1台 • 入库服务节点:24台 • Web查询应用服务节点:20台 • 机架间通过万兆交换机连接 • 网络冗余 产创平台 采集系统 核心交换机 核心交换机 负载均衡器 负载均衡器 防火墙 防火墙 DCN网络接入 路由器 核心交换机 核心交换机 防火墙 防火墙 接入路由器 查询接口服务器 DataNode and RegionServer 核心交换机 核心交换机 FTP传输及入库 NameNode节点 Zookeeper X 省 客服系统 GGSN SGSN Gn口 产创平台 采集系统0 码力 | 45 页 | 4.33 MB | 1 年前3
HBase基本介绍• NameNode/DataNode 系统组成 如图有这么⼏几个组成部分, 前两个是HBase的 Master是负责管理理的, RegionServer是实际⼲干活的 zookeeper作为协调信息存储的地⽅方, ⽐比如节点健康状态 然后HBase的数据都要存放在hdfs上, 就要有node. 如图可以看出RegionServer和Datanode尽量量在同⼀一台机器器上.0 码力 | 33 页 | 4.86 MB | 1 年前3
HBase Practice At XiaoMifailure since each datanode need to handle too many streams. Number of streams = number of Log * number of region * number of replica * number of regionserver Streams each datanode need to handle = Number0 码力 | 56 页 | 350.38 KB | 1 年前3
Apache ShardingSphere 5.2.0 DocumentRESOURCES(resource [, resource] ...) dataNodes: DATANODES(dataNode [, dataNode] ...) resource: resourceName | inlineExpression dataNode: resourceName | inlineExpression shardingColumn: SHARDING_COLUMN=columnName tableDefinition | autoTableDefinition ) )* tableDefinition ::= tableName '(' 'DATANODES' '(' dataNode ( ',' dataNode )* ')' ( ',' 'DATABASE_ STRATEGY' '(' strategyDefinition ')' )? ( ',' 'TABLE_STRATEGY' '(' tableDefinition | autoTableDefinition ) )* tableDefinition ::= tableName '(' 'DATANODES' '(' dataNode ( ',' dataNode )* ')' ( ',' 'DATABASE_ STRATEGY' '(' strategyDefinition ')' )? ( ',' 'TABLE_STRATEGY' '('0 码力 | 483 页 | 4.27 MB | 1 年前3
Apache ShardingSphere 5.2.1 DocumentRESOURCES(resource [, resource] ...) dataNodes: DATANODES(dataNode [, dataNode] ...) resource: resourceName | inlineExpression dataNode: dataNodeName | inlineExpression shardingColumn: SHARDING_COLUMN=columnName tableDefinition | autoTableDefinition ) )* tableDefinition ::= tableName '(' 'DATANODES' '(' dataNode ( ',' dataNode )* ')' ( ',' 'DATABASE_ STRATEGY' '(' strategyDefinition ')' )? ( ',' 'TABLE_STRATEGY' '(' tableDefinition | autoTableDefinition ) )* tableDefinition ::= tableName '(' 'DATANODES' '(' dataNode ( ',' dataNode )* ')' ( ',' 'DATABASE_ STRATEGY' '(' strategyDefinition ')' )? ( ',' 'TABLE_STRATEGY' '('0 码力 | 523 页 | 4.51 MB | 1 年前3
Apache ShardingSphere 中文文档 5.2.0RESOURCES(resource [, resource] ...) dataNodes: DATANODES(dataNode [, dataNode] ...) resource: resourceName | inlineExpression dataNode: resourceName | inlineExpression shardingColumn: SHARDING_COLUMN=columnName tableDefinition | autoTableDefinition ) )* tableDefinition ::= tableName '(' 'DATANODES' '(' dataNode ( ',' dataNode )* ')' ( ',' 'DATABASE_ STRATEGY' '(' strategyDefinition ')' )? ( ',' 'TABLE_STRATEGY' '(' tableDefinition | autoTableDefinition ) )* tableDefinition ::= tableName '(' 'DATANODES' '(' dataNode ( ',' dataNode )* ')' ( ',' 'DATABASE_ STRATEGY' '(' strategyDefinition ')' )? ( ',' 'TABLE_STRATEGY' '('0 码力 | 449 页 | 5.85 MB | 1 年前3
Apache ShardingSphere 5.4.1 DocumentifNotExists ::= 'IF' 'NOT' 'EXISTS' tableDefinition ::= tableName '(' 'DATANODES' '(' dataNode (',' dataNode)* ')' (',' 'DATABASE_ STRATEGY' '(' strategyDefinition ')')? (',' 'TABLE_STRATEGY' '(' value (',' key '=' value)* ')' key ::= string value ::= literal tableName ::= identifier dataNode ::= string storageUnitName ::= identifier columnName ::= identifier algorithmType ::= identifier (tableDefinition | autoTableDefinition))* tableDefinition ::= tableName '(' 'DATANODES' '(' dataNode (',' dataNode)* ')' (',' 'DATABASE_ STRATEGY' '(' strategyDefinition ')')? (',' 'TABLE_STRATEGY' '('0 码力 | 572 页 | 3.73 MB | 1 年前3
Apache ShardingSphere 中文文档 5.4.1ifNotExists ::= 'IF' 'NOT' 'EXISTS' tableDefinition ::= tableName '(' 'DATANODES' '(' dataNode (',' dataNode)* ')' (',' 'DATABASE_ STRATEGY' '(' strategyDefinition ')')? (',' 'TABLE_STRATEGY' '(' value (',' key '=' value)* ')' key ::= string value ::= literal tableName ::= identifier dataNode ::= string storageUnitName ::= identifier columnName ::= identifier algorithmType ::= identifier (tableDefinition | autoTableDefinition))* tableDefinition ::= tableName '(' 'DATANODES' '(' dataNode (',' dataNode)* ')' (',' 'DATABASE_ STRATEGY' '(' strategyDefinition ')')? (',' 'TABLE_STRATEGY' '('0 码力 | 530 页 | 4.49 MB | 1 年前3
共 20 条
- 1
- 2













