Python、C、Perl、TCL、 PLSQL 等等语言来扩展功能，在后续章节中，我将展现这种扩展 是如何的方便，另外，开发新的功能模块、新的数据类型、新的索 引类型等等非常方便，只要按照 API 接口开发，无需对 PG 重新编译。 PG 中 contrib 目录下的各个第三方模块，在 GP 中的 postgis 空间 数据库、R、Madlib、pgcrypto 各类加密算法、gptext 全文检索都 是通过这种方式实现功能扩展的。 postgresql 的 contrib 中的一些常用模块 Greenplum 提 供了编译后的模块开箱即用，如：oraface、postgis、pgcrypt 等， 对于其它模块，用户可以自行将 contrib 下的代码与 Greenplum 的 include 头文件编译后，将动态 so 库文件部署到所有节点就可进行测 试使用了。有些模块还是非常好用的，例如：oraface，基本上集成了 Oracle Hadoop 集群配合运作的案例越 来越多。在中国移动的大数据架 构规范中，云化 ETL 是一个重 要的组成部分。云化 ETL 就是构架在 Hadoop 集群之上。Greenplum 提供了专用产品模块 gphdfs，Greenplum 通过 gphdfs 可以直接与 HDFS 上的数据进行交互，并且可以同时发挥 Greenplum 和 Hadoop 两者并行处理的优势。 Big Date2.indd

Client、Master Node 和 Compute Node 组成。基本组成部分的功能如下： 产品架构 Greenplum数据仓库 UDW Copyright © 2012-2021 UCloud 优刻得 7/206 1. Client：访问 UDW 的客⼾端 ⽀持通过 JDBC、ODBC、PHP、Python、命令⾏ Sql 等⽅式访问 UDW 2. Master Node：访问 UDW 数据仓库的⼊⼝ Compute Node： Compute Node 管理节点的计算和存储资源 每个 Compute Node 由多个 Segment 组成 Segment 负责业务数据的存储、⽤⼾ SQL 的执⾏ ⾼可⽤ ⾼可⽤ 产品架构 Greenplum数据仓库 UDW Copyright © 2012-2021 UCloud 优刻得 8/206 如上图所⽰： 1. Compute Node 中任⼀ Segment Segment 都会有⼀个 Mirror Segment 备份到其他的 Compute Node 上，当 Primary Segment 出现不可⽤的时候会⾃动切换到 Mirror Segment， 当 Primary Segment 恢复之后，Primary Segment 会⾃动恢复这期间的变更。 2. Master 节点是主从模式，当 Active Master 不可⽤时会⾃动切换到 Standby

数据仓库体系架构 一 Greenplum开发规范 五 Greenplum运维体系 四 Greenplum扩展规划 六 39 Greenplum扩展规划 整体扩展思路 • OLAP三大模块不断完善 Ø 通过dbsync程序，实现数据传输和加载优化 Ø 对于Greenplum上的架构和设计不断优化 Ø 继续建设多样化的postgresql数据集市，满足不同需求 • 优化现有业务的调度实现

bool delayEagerFree 某个Segment的排序节点输出最后一条元组 后是否可以提前 释放内存 36 ● ExecSort： 从下层Outer节点读取所有元组，并传递给tuplesort模块进行排序 排序节点 37 TupleSort是排序节点的核心，算法主要阶段： ● 第一阶段初始化TupleSort，通过调用函数tuplesort_begin_common，生成 Tuple

的支持。这些新数据类型及相关函数有助于打造 一个没有孤岛的环境，结构化、半结构化和非结构化数据均可在一个位置进行存储和分析。其他功能还有 dblink，这是一 个用于与其他数据库（无论在同一数据库主机还是远程主机上）建立连接的新模块。Pivotal Greenplum 支持数据库用户 使用 dblink 在其他数据库中执行简短的临时查询。最后，Greenplum 5 将已安装的 Python 版本升级为 2.7。PL/Python

kube-proxy docker Node kubelet kube-proxy docker Node Greenplum on Kubernetes 101 Kubernetes Master Pod kubelet kube-proxy docker Node Pod kubelet kube-proxy docker Node Greenplum on Kubernetes kube-proxy docker Node Pod kubelet kube-proxy docker Node Storage volumes Greenplum on Kubernetes 101 Kubernetes Master Pod Postgres Container kubelet kube-proxy docker Node Pod Postgres Container kubelet kube-proxy docker Node Storage volumes Greenplum on Kubernetes 101 Kubernetes Master Greenplum Service Pod Postgres container kubelet kube-proxy docker Node Pod Postgres Container

subquery with an explicit sequence nextval. 178253995 Resolved an issue where the sequence executor node was generating duplicate sequence values. Upgrading from Greenplum 6.x to Greenplum 6.17 See Upgrading Stack Regardless of the hardware platform you choose, a production Greenplum Database processing node (a segment host) is typically configured as described in this section. The segment hosts do the majority following to avoid this issue. When you add the gpadmin operating system user account to the master node in Creating the Greenplum Administrative User, create the user as a system account. Disable RemoveIPC

variables on the query executor. 12161 : Resolved an issue with dispatch conditions and redundant motion node that was generating an assert failure in execMotionUnsortedReceiver. 12648 : Resolved an issue where memory. 31734 : Greenplum was generating the error ERROR: interconnect error: A HTAB entry for motion node 77 already exists after creating the extension for the ltree module. This issue has been resolved alternative locations than the ones expected. 12419 : Resolves an issue where Greenplum Database, during node recovery, generated an error when the field standby_mode=on was set in the recovery.conffile. The

159 31734 : Greenplum was generating the error ERROR: interconnect error: A HTAB entry for motion node 77 already exists after creating the extension for the ltree module. This issue has been resolved alternative locations than the ones expected. 12419 : Resolves an issue where Greenplum Database, during node recovery, generated an error when the field standby_mode=on was set in the recovery.conffile. The Query Optimizer 31733 : Queries were crashing due to GPORCA prematurely terminating the motion node before interconnect was torn down. This issue is now resolved. 31640 : Running queries with GPORCA

32770 Resolves an issue where a query that produced a Left Anti Semi Join (LASJ) (NOT IN) plan node was gathered and run on the Greenplum Database master host, generating many spill files and ultimately ensuring that squelching mechanics are applied after a SIGPIPE for queries that excecute on the master node, as they are applied for queries executed on segment nodes. Query Processing 32481 Resolved an login attempts are no longer a security risk. 13860 The EPQ routine for AO/AOCS relations and DML node has been disallowed. 13699 Duplicate gp_fastsequence values (ctid) no longer appear in AO tables