Greenplum tablespace与filespace的关系

GP里面和PG不同的是，多了个filespace概念，在GP里面tablespace必须建立在filespace上面，所以建立表空间之前，先要有指定的filespace

filespace可以认为是多个文件目录的集合，可以对应一个或者多个tablespace

下面介绍一下创建filespace的实例：

1. su - gpadmin

2. 创建新的目录赋予权限

192-168-101-115

192-168-101-116

mkdir -p /data01/gpadmin/gpdata/master_hank

192-168-100-217

192-168-100-225

mkdir -p /data01/gpadmin/gpdata/primary_hank

mkdir -p /data01/gpadmin/gpdata/mirror_hank

赋予权限

chown -R gpadmin.gpadmin /data01/gpadmin/gpdata/master_hank /data01/gpadmin/gpdata/primary_hank /data01/gpadmin/gpdata/mirror_hank

3.生成配置文件

gpfilespace -o filespace_hank

20170317:17:05:13:020913 gpfilespace:db-192-168-101-115:gpadmin-[INFO]:-getting config

Enter a name for this filespace

> hank

Checking your configuration:

Your system has 2 hosts with 1 primary and 1 mirror segments per host.

Your system has 2 hosts with 0 primary and 0 mirror segments per host.

Configuring hosts: [db-192-168-100-225.sky-mobi.com, db-192-168-100-217.sky-mobi.com]

Please specify 1 locations for the primary segments, one per line:

primary location 1> /data01/gpadmin/gpdata/primary_hank

Please specify 1 locations for the mirror segments, one per line:

mirror location 1> /data01/gpadmin/gpdata/mirror_hank

Configuring hosts: [db-192-168-101-116.sky-mobi.com, db-192-168-101-115.sky-mobi.com]

Enter a file system location for the master

master location> /data01/gpadmin/gpdata/master_hank

20170317:17:05:34:020913 gpfilespace:db-192-168-101-115:gpadmin-[INFO]:-Creating configuration file...

20170317:17:05:34:020913 gpfilespace:db-192-168-101-115:gpadmin-[INFO]:-[created]

20170317:17:05:34:020913 gpfilespace:db-192-168-101-115:gpadmin-[INFO]:-

To add this filespace to the database please run the command:

   gpfilespace --config /home/gpadmin/filespace_hank

创建filespace

gpfilespace -c filespace_hank 

20170317:17:06:09:021026 gpfilespace:db-192-168-101-115:gpadmin-[INFO]:-

A tablespace requires a file system location to store its database

files. A filespace is a collection of file system locations for all components

in a Greenplum system (primary segment, mirror segment and master instances).

Once a filespace is created, it can be used by one or more tablespaces.

20170317:17:06:09:021026 gpfilespace:db-192-168-101-115:gpadmin-[INFO]:-getting config

Reading Configuration file: 'filespace_hank'

20170317:17:06:09:021026 gpfilespace:db-192-168-101-115:gpadmin-[INFO]:-Performing validation on paths

..............................................................................

20170317:17:06:10:021026 gpfilespace:db-192-168-101-115:gpadmin-[INFO]:-Connecting to database

20170317:17:06:10:021026 gpfilespace:db-192-168-101-115:gpadmin-[INFO]:-Filespace "hank" successfully created

其实这里就是执行的create filespace语句

CREATE FILESPACE hank

( 

  1: '/data01/gpadmin/gpdata/master_hank/gpseg-1',

  2: '/data01/gpadmin/gpdata/primary_hank/gpseg0',

  3: '/data01/gpadmin/gpdata/primary_hank/gpseg1',

  4: '/data01/gpadmin/gpdata/mirror_hank/gpseg0',

  5: '/data01/gpadmin/gpdata/mirror_hank/gpseg1',

  6: '/data01/gpadmin/gpdata/master_hank/gpseg-1'

);

postgres=# select * from pg_filespace;

  fsname   | fsowner 

-----------+---------

 pg_system |      10

 hank      |      10

postgres=# create TABLESPACE tbs_hank FILESPACE hank;

CREATE TABLESPACE

SELECT spcname as tblspc, fsname as filespc,fsedbid as seg_dbid, fselocation as datadir

FROM pg_tablespace pgts, pg_filespace pgfs,

pg_filespace_entry pgfse

WHERE pgts.spcfsoid=pgfse.fsefsoid

AND pgfse.fsefsoid=pgfs.oid

ORDER BY tblspc, seg_dbid;

   tblspc   |  filespc  | seg_dbid |                  datadir                   

------------+-----------+----------+--------------------------------------------

 pg_default | pg_system |        1 | /data01/gpadmin/gpdata/master/gpseg-1

 pg_default | pg_system |        2 | /data01/gpadmin/gpdata/primary/gpseg0

 pg_default | pg_system |        3 | /data01/gpadmin/gpdata/primary/gpseg1

 pg_default | pg_system |        4 | /data01/gpadmin/gpdata/mirror/gpseg0

 pg_default | pg_system |        5 | /data01/gpadmin/gpdata/mirror/gpseg1

 pg_default | pg_system |        6 | /data01/gpadmin/gpdata/master/gpseg-1

 pg_global  | pg_system |        1 | /data01/gpadmin/gpdata/master/gpseg-1

 pg_global  | pg_system |        2 | /data01/gpadmin/gpdata/primary/gpseg0

 pg_global  | pg_system |        3 | /data01/gpadmin/gpdata/primary/gpseg1

 pg_global  | pg_system |        4 | /data01/gpadmin/gpdata/mirror/gpseg0

 pg_global  | pg_system |        5 | /data01/gpadmin/gpdata/mirror/gpseg1

 pg_global  | pg_system |        6 | /data01/gpadmin/gpdata/master/gpseg-1

 tbs_hank   | hank      |        1 | /data01/gpadmin/gpdata/master_hank/gpseg-1

 tbs_hank   | hank      |        2 | /data01/gpadmin/gpdata/primary_hank/gpseg0

 tbs_hank   | hank      |        3 | /data01/gpadmin/gpdata/primary_hank/gpseg1

 tbs_hank   | hank      |        4 | /data01/gpadmin/gpdata/mirror_hank/gpseg0

 tbs_hank   | hank      |        5 | /data01/gpadmin/gpdata/mirror_hank/gpseg1

 tbs_hank   | hank      |        6 | /data01/gpadmin/gpdata/master_hank/gpseg-1

建库，赋予权限，建表实例：

CREATE ROLE hank LOGIN  ENCRYPTED PASSWORD 'hank'  nosuperuser noinherit nocreatedb nocreaterole ;

CREATE DATABASE hank

WITH  OWNER = gpadmin 

      TEMPLATE = template0 

      ENCODING = 'UTF8' 

      TABLESPACE = tbs_hank;

grant all on database hank to hank   with grant option;

grant all on tablespace tbs_hank to hank; 

\c  hank hank 

create schema hank;

grant usage on schema hank to hank;

create table tb1 (a bigint,b text) tablespace tbs_hank distributed by (a);

insert into tb1 select generate_series(1,10000),'hank'||generate_series(1,10000);