Greenplum数据库基础培训.ppt

Greenplum数据库 海量并行处理 (Massively Parallel Processing) DBMS 基于 PostgreSQL 8.2 相同的客户端功能 增加支持并行处理的技术 增加支持数据仓库和BI的特性 外部表(external tables)/并行加载(parallel loading) 资源管理 查询优化器增强(query optimizer enhancements) S1002 Network Configuration Greenplum 基本体系架构 客户端接口和程序 Master Host 函数 字符串处理函数 Substr/length/lower/upper/trim/replace/position rPad/lpad To_char, || (字符串连接) substring(string from pattern ～/ ～* ，like，simillar to (模式匹配) 其它杂类 Case 。。。When/Coalesce nullif generate_series In/not in/exists/any/all QA 问题与解答 * Instructor Notes By default, the master host is configured to allow remote client connections from the default superuser account only. This is the user who initialized the Greenplum array. In this case, the OS user name and the Greenplum database user name are the same, but this may not always be the case for other users. Let’s go over each line of the pg_hba.conf file for the master host and explain what is happening: Lines 1 and 2 are saying that for local UNIX domain socket connections allow the gpadmin user (the superuser who bootstrapped the array, lab# in our training environment) and all users to connect to all databases. The ident authentication method works by obtaining the clients OS user name, then determining the allowed database user names using a map file. There is a predefined ident map named sameuser, which allows any operating system user to connect as the database user of the same name. Any other maps must be created manually using the pg_ident.conf file. Line 3 is saying that for TCP/IP connections allow only the gpadmin superuser to connect to all databases when the connection comes from the local host. Note that the CIDR address is in IPv4 format. trust means no authentication is performed. Line 4 is saying that for any remote host TCP/IP connections allow only the gpadmin superuser to connect to all databases. Note that the CIDR address is in IPv4 format. md5 means password authentication is performed. If no password is provide