GREENPLUM总体结构：

 

 

 

数据库由Master Severs和Segment Severs通过Interconnect互联组成。

Master主机负责：建立与客户端的连接和管理；SQL的解析并形成执行计划；执行计划向Segment的分发收集Segment的执行结果；Master不存储业务数据，只存储数据字典。 

Segment主机负责：业务数据的存储和存取；用户查询SQL的执行。

master node高可用，类似于hadoop的namenode和second namenode，实现主备的高可用。

Greenplum使用MPP架构：

 

 

 

1、准备

　　这里准备了4台服务器，1台做master，1台做standby，4台都做存储

 

OS: Red Hat Enterprise Linux Server release 6.5 (Santiago)

GP: greenplum-db-4.3.8.2-build-1-RHEL5-x86_64.zip

JDK:1.8

192.168.244.110 node01                 #master node

192.168.244.111 node02                 #standby node

192.168.244.112 node03                 #segment node

192.168.244.113 node04                 #segment node

 

 

2、系统参数配置

　　系统参数配置的修改需要在每个节点服务器上执行

2.1.修改Linux内核参数

 

[root@node01 ~]# vi /etc/sysctl.conf

# Kernel sysctl configuration file for Red Hat Linux

#

# For binary values, 0 is disabled, 1 is enabled.  See sysctl(8) and

# sysctl.conf(5) for more details.

#

# Use '/sbin/sysctl -a' to list all possible parameters.

 

# Controls IP packet forwarding

net.ipv4.ip_forward = 0

 

# Controls source route verification

net.ipv4.conf.default.rp_filter = 1

 

# Do not accept source routing

net.ipv4.conf.default.accept_source_route = 0

 

# Controls the System Request debugging functionality of the kernel

kernel.sysrq = 1

 

# Controls whether core dumps will append the PID to the core filename.

# Useful for debugging multi-threaded applications.

kernel.core_uses_pid = 1

 

# Controls the use of TCP syncookies

net.ipv4.tcp_syncookies = 1

 

# Disable netfilter on bridges

#net.bridge.bridge-nf-call-ip6tables = 0

#net.bridge.bridge-nf-call-iptables = 0

#net.bridge.bridge-nf-call-arptables = 0

 

# Controls the default maxmimum size of a mesage queue

kernel.msgmnb = 65536

 

# Controls the maximum size of a message, in bytes

kernel.msgmax = 65536

kernel.msgmni = 2048

 

kernel.sem = 250 512000 100 2048

 

# Controls the maximum shared segment size, in bytes

#kernel.shmmax = 68719476736

kernel.shmmax = 500000000

kernel.shmmni = 4096

 

# Controls the maximum number of shared memory segments, in pages

#kernel.shmall = 4294967296

kernel.shmall = 4000000000

 

net.ipv4.tcp_tw_recycle=1

net.ipv4.tcp_max_syn_backlog=4096

net.ipv4.ip_local_port_range = 1025 65535

net.core.netdev_max_backlog=10000

vm.overcommit_memory=2

net.ipv4.conf.all.arp_filter = 1

net.core.rmem_max = 2097152

net.core.wmem_max = 2097152

 

 

2.2.修改Linux最大限制

 

[root@node01 ~]# vi /etc/security/limits.conf

#greenplum configs

* soft nofile 65536

* hard nofile 65536

* soft nproc 131072

* hard nproc 131072

 

 

2.3.I/O调整优化

[root@node01 ~]# vi /boot/grub/menu.lst

#greenplum configs

elevator=deadline

 

2.4.添加所有节点到HOST

[root@node01 ~]# vi /etc/hosts

192.168.244.110 node01

192.168.244.111 node02

192.168.244.112 node03

192.168.244.113 node04

 

2.5.关闭防火墙

[root@node01 ~]# chkconfig --list iptables

[root@node01 ~]# chkconfig --level 0123456 iptables off

 

2.6.设置SELINUX

[root@node01 ~]# vi /etc/selinux/config

SELINUX=disabled

 

2.7.重启系统使配置生效

[root@node01 ~]# reboot

 

3、GP安装

　　GP的安装操作都是在主节点master上执行的

3.1.创建gpadmin用户

所有节点创建gpadmin用户

[root@node01 ~]# useradd gpadmin

[root@node01 ~]# passwd gpadmin

 

3.2.设置gpadmin用户环境

 

[gpadmin@node01 ~]$ cd /home/gpadmin

[gpadmin@node01 ~]$ vi .bashrc

[gpadmin@node01 ~]$ vi .bash_profile

.bashrc和.bash_profile最后都添加下面两行

source /usr/local/greenplum-db/greenplum_path.sh
export MASTER_DATA_DIRECTORY=/data1/gpdata/master/gpseg-1

 

/etc/profile也添加上面两行内容

 

设置完后记得source一下使其立即生效

 

3.3.上传并解压安装包

将greenplum-db-4.3.8.2-build-1-RHEL5-x86_64.zip上传至master的/opt/目录下

[root@node01 opt]# unzip greenplum-db-4.3.8.2-build-1-RHEL5-x86_64.zip

[root@node01 opt]#/bin/bash greenplum-db-4.3.8.2-build-1-RHEL5-x86_64.bin

按提示输入回车或yes 这一步会将安装包解压到/usr/local/下，并建立软连接greenplum-db

 

 

3.4.准备节点服务器信息文件

后面的批量安装会用到这两个文件，如果all_host和all_segment内容一样，可以只创建一个文件

 

[root@node01 greenplum-db]# mkdir -p /opt/gpinst/ [root@node01 gpinst]# touch all_host

[root@node01 gpinst]# touch all_segment

all_host和all_segment内容：

node01

node02

node03

node04

 

3.5.建立节点服务器间的信任

[root@node01 gpinst]# gpssh-exkeys -f /opt/gpinst/all_host

[STEP 1 of 5] create local ID and authorize on local host

 

[STEP 2 of 5] keyscan all hosts and update known_hosts file

 

[STEP 3 of 5] authorize current user on remote hosts

  ... send to node02

  ***

  *** Enter password for node02:

  ... send to node03

  ... send to node04

 

[STEP 4 of 5] determine common authentication file content

 

[STEP 5 of 5] copy authentication files to all remote hosts

  ... finished key exchange with node02

  ... finished key exchange with node03

  ... finished key exchange with node04

 

[INFO] completed successfully

按照提示输入root密码，记住这一步不能输入gpadmin的密码，因为批量安装时需要在/usr/local下创建目录，需要root权限

 

3.6.批量安装

[root@node01 gpinst]# gpseginstall -f /opt/gpinst/all_host -u gpadmin -p gpadmin

20170717:13:13:51:026424 gpseginstall:node01:root-[INFO]:-Installation Info:

link_name greenplum-db

binary_path /usr/local/greenplum-db-4.3.8.2

binary_dir_location /usr/local

binary_dir_name greenplum-db-4.3.8.2

20170717:13:13:51:026424 gpseginstall:node01:root-[INFO]:-check cluster password access

。。。。。。

 

 

20170717:13:16:39:026424 gpseginstall:node01:root-[INFO]:-SUCCESS -- Requested commands completed

 

这一步其实就是将master上的greenplum打包通过scp命令传到all_host中的主机上，并赋予目录gpadmin的权限

 

3.7.检查批量安装情况

[root@node01 local]# gpssh -f /opt/gpinit/all_host -e ls -l $GPHOME

[root@node01 gpinst]# gpssh -f /opt/gpinst/all_host -e ls -l $GPHOME

[node01] ls -l /usr/local/greenplum-db/.

[node01] total 276

[node01] drwxr-xr-x 4 gpadmin gpadmin   4096 May 10  2016 bin

[node01] drwxr-xr-x 2 gpadmin gpadmin   4096 May 10  2016 demo

[node01] drwxr-xr-x 5 gpadmin gpadmin   4096 May 10  2016 docs

[node01] drwxr-xr-x 2 gpadmin gpadmin   4096 May 10  2016 etc

[node01] drwxr-xr-x 3 gpadmin gpadmin   4096 May 10  2016 ext

[node01] -rw-r--r-- 1 gpadmin gpadmin  43025 May 10  2016 GPDB-LICENSE.txt

[node01] -rw-r--r-- 1 gpadmin gpadmin    735 Jul 17 13:06 greenplum_path.sh

[node01] drwxr-xr-x 6 gpadmin gpadmin   4096 May 10  2016 include

[node01] drwxr-xr-x 9 gpadmin gpadmin   4096 May 10  2016 lib

[node01] -rw-r--r-- 1 gpadmin gpadmin 192912 May 10  2016 LICENSE.thirdparty

[node01] drwxr-xr-x 2 gpadmin gpadmin   4096 May 10  2016 sbin

[node01] drwxr-xr-x 4 gpadmin gpadmin   4096 May 10  2016 share

[node02] ls -l /usr/local/greenplum-db/.

[node02] total 276

[node02] drwxr-xr-x 4 gpadmin gpadmin   4096 May 10  2016 bin

[node02] drwxr-xr-x 2 gpadmin gpadmin   4096 May 10  2016 demo

[node02] drwxr-xr-x 5 gpadmin gpadmin   4096 May 10  2016 docs

返回结果中各节点目录一致则成功

 

3.8.创建存储目录

 

master

[root@node01 gpinst]# mkdir -p /data/gpdata/master

[root@node01 gpinst]# chown gpadmin:gpadmin /data/gpdata/master

 

segment

[root@node01 gpinst]# gpssh -f /opt/gpinst/all_host -e 'mkdir -p /data/gpdata/primary'

[node04] mkdir -p /data/gpdata/primary

[node02] mkdir -p /data/gpdata/primary

[node01] mkdir -p /data/gpdata/primary

[node03] mkdir -p /data/gpdata/primary

 

[root@node01 gpinst]# gpssh -f /opt/gpinst/all_host -e 'chown gpadmin:gpadmin /data/gpdata/primary'

 

[node01] chown gpadmin:gpadmin /data/gpdata/primary

[node02] chown gpadmin:gpadmin /data/gpdata/primary

[node04] chown gpadmin:gpadmin /data/gpdata/primary

[node03] chown gpadmin:gpadmin /data/gpdata/primary

mirror
[root@node01 local]# gpssh -f /opt/gpinit/all_segment -e 'mkdir -p /data1/gpdata/mirror'
[root@node01 local]# gpssh -f /opt/gpinit/all_segment -e 'chown gpadmin:gpadmin /data1/gpdata/mirror'

 

 

 

#3.9.设置时钟同步

 

vi /etc/ntp.conf 在server第一行添加下面两行
server 192.168.244.110

重启ntpd服务 /etc/init.d/ntpd restart

查看ntp同步情况 ntpq -p

使ntpd服务重启服务器后也启动 chkconfig --level 0123456 ntpd on

 

 

3.10.创建并修改GP初始化文件

[root@node01 gpadmin]# su - gpadmin

[gpadmin@node01 ~]$ mkdir /home/gpadmin/gpconfigs

[gpadmin@node01 ~]$ ll

total 4

drwxrwxr-x 2 gpadmin gpadmin 4096 Jul 17 13:30 gpconfigs

[gpadmin@node01 ~]$ cp /usr/local/greenplum-db/docs/cli_help/gpconfigs/gpinitsystem_config /home/gpadmin/gpconfigs/

[gpadmin@node01 ~]$ chmod 775 /home/gpadmin/gpconfigs/gpinitsystem_config

 

修改GP初始化文件

 

[gpadmin@node01 ~]$ vi gpconfigs/gpinitsystem_config

 

declare -a DATA_DIRECTORY=(/data/gpdata/primary /data/gpdata/primary /data1/prim# FILE NAME: gpinitsystem_config

 

# Configuration file needed by the gpinitsystem

 

################################################

#### REQUIRED PARAMETERS

################################################

 

#### Name of this Greenplum system enclosed in quotes.

ARRAY_NAME="HX Greenplum DW"

 

#### Naming convention for utility-generated data directories.

SEG_PREFIX=gpseg

 

#### Base number by which primary segment port numbers

#### are calculated.

PORT_BASE=40000

 

#### File system location(s) where primary segment data directories

#### will be created. The number of locations in the list dictate

#### the number of primary segments that will get created per

#### physical host (if multiple addresses for a host are listed in

#### the hostfile, the number of segments will be spread evenly across

#### the specified interface addresses).

declare -a DATA_DIRECTORY=(/data/gpdata/primary /data/gpdata/primary)

 

#### OS-configured hostname or IP address of the master host.

MASTER_HOSTNAME=node01

 

#### File system location where the master data directory

#### will be created.

MASTER_DIRECTORY=/data/gpdata/master

 

#### Port number for the master instance.

MASTER_PORT=5432

 

#### Shell utility used to connect to remote hosts.

TRUSTED_SHELL=ssh

 

#### Maximum log file segments between automatic WAL checkpoints.

CHECK_POINT_SEGMENTS=8

 

#### Default server-side character set encoding.

# FILE NAME: gpinitsystem_config

 

# Configuration file needed by the gpinitsystem

 

################################################

#### REQUIRED PARAMETERS

################################################

 

#### Name of this Greenplum system enclosed in quotes.

ARRAY_NAME="HX Greenplum DW"

 

#### Naming convention for utility-generated data directories.

SEG_PREFIX=gpseg

 

#### Base number by which primary segment port numbers

#### are calculated.

PORT_BASE=40000

 

#### File system location(s) where primary segment data directories

#### will be created. The number of locations in the list dictate

#### the number of primary segments that will get created per

#### physical host (if multiple addresses for a host are listed in

#### the hostfile, the number of segments will be spread evenly across

#### the specified interface addresses).

declare -a DATA_DIRECTORY=(/data/gpdata/primary /data/gpdata/primary)

 

#### OS-configured hostname or IP address of the master host.

MASTER_HOSTNAME=node01

 

#### File system location where the master data directory

#### will be created.

MASTER_DIRECTORY=/data/gpdata/master

 

#### Port number for the master instance.

MASTER_PORT=5432

 

#### Shell utility used to connect to remote hosts.

TRUSTED_SHELL=ssh

 

#### Maximum log file segments between automatic WAL checkpoints.

CHECK_POINT_SEGMENTS=8

 

#### Default server-side character set encoding.

ENCODING=UNICODE

 

################################################

#### OPTIONAL MIRROR PARAMETERS

################################################

 

#### Base number by which mirror segment port numbers

#### are calculated.

MIRROR_PORT_BASE=50000

 

#### Base number by which primary file replication port

#### numbers are calculated.

REPLICATION_PORT_BASE=41000

 

#### Base number by which mirror file replication port

#### numbers are calculated.

MIRROR_REPLICATION_PORT_BASE=51000

 

#### File system location(s) where mirror segment data directories

#### will be created. The number of mirror locations must equal the

#### number of primary locations as specified in the

#### DATA_DIRECTORY parameter.

declare -a MIRROR_DATA_DIRECTORY=(/data/gpdata/mirror /data/gpdata/mirror)

 

 

################################################

#### OTHER OPTIONAL PARAMETERS

################################################

 

#### Create a database of this name after initialization.

DATABASE_NAME=hx_gp

 

#### Specify the location of the host address file here instead of

#### with the the -h option of gpinitsystem.

MACHINE_LIST_FILE=/home/gpadmin/gpconfigs/hostfile_gpinitsystem

"gpconfigs/gpinitsystem_config" 78L, 2560C written     

 

3.11.初始化GP

[gpadmin@node01 ~]$ gpinitsystem -c /home/gpadmin/gpconfigs/gpinitsystem_config  -h /opt/gpinst/all_host

 

20170717:13:37:48:027877 gpinitsystem:node01:gpadmin-[INFO]:-Checking configuration parameters, please wait...

20170717:13:37:48:027877 gpinitsystem:node01:gpadmin-[INFO]:-Reading Greenplum configuration file /home/gpadmin/gpconfigs/gpinitsystem_config

20170717:13:37:48:027877 gpinitsystem:node01:gpadmin-[INFO]:-Locale has not been set in /home/gpadmin/gpconfigs/gpinitsystem_config, will set to default value

20170717:13:37:48:027877 gpinitsystem:node01:gpadmin-[INFO]:-Locale set to en_US.utf8

20170717:13:37:49:027877 gpinitsystem:node01:gpadmin-[INFO]:-MASTER_MAX_CONNECT not set, will set to default value 250

20170717:13:37:49:027877 gpinitsystem:node01:gpadmin-[INFO]:-Checking configuration parameters, Completed

20170717:13:37:49:027877 gpinitsystem:node01:gpadmin-[INFO]:-Commencing multi-home checks, please wait...

中间需要输入一次：Y

Continue with Greenplum creation Yy/Nn>

y

20170717:13:38:24:027877 gpinitsystem:node01:gpadmin-[INFO]:-Building the Master instance database, please wait...

如果没有报ERROR，GP数据库就安装好了。 

查看数据库名称：hx_gp

 [gpadmin@node01 ~]$ psql -l

                  List of databases

   Name    |  Owner  | Encoding |  Access privileges 

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

 hx_gp     | gpadmin | UTF8     |

 postgres  | gpadmin | UTF8     |

 template0 | gpadmin | UTF8     | =c/gpadmin         

                                : gpadmin=CTc/gpadmin

 template1 | gpadmin | UTF8     | =c/gpadmin         

                                : gpadmin=CTc/gpadmin

(4 rows)

 

3.12 增加standby

一个gp集群只有一个master肯定会让人不放心，还好有备用，当master宕掉后，会自动启用standby作为master

 

在standby服务器上执行
[root@node02 ~]# mkdir /data1/gpdata/master

[root@node02 ~]# chown gpadmin:gpadmin /data1/gpdata/master

在master服务器上执行 
[gpadmin@node01 ~]$ gpinitstandby -s node02
中间输入一次Y

20170717:16:29:14:025010 gpinitstandby:node01:gpadmin-[INFO]:-Validating environment and parameters for standby initialization...

20170717:16:29:14:025010 gpinitstandby:node01:gpadmin-[INFO]:-Checking for filespace directory /data/gpdata/master/gpseg-1 on node02

20170717:16:29:14:025010 gpinitstandby:node01:gpadmin-[INFO]:------------------------------------------------------

。。。。。。

。。。。。。

20170717:16:29:31:025010 gpinitstandby:node01:gpadmin-[INFO]:-Backup files of pg_hba.conf cleaned up successfully.

20170717:16:29:31:025010 gpinitstandby:node01:gpadmin-[INFO]:-Successfully created standby master on node02

 

 查看状态：gpstate

 

3.13增加mirror

mirror就是镜像，也叫数据备份。mirror对于数据存储来说很重要，因为我们的服务器指不定什么时候出毛病，有mirror就好很多了，因为两台存储节点同时宕掉的几率还是很小的。如果前面在GP初始化文件里忘记配置mirror了，请按照下面的方法添加

[gpadmin@node01 ~]$ gpaddmirrors -p 1000
运行过程中需要输入两次mirror路径：/data1/gpdata/mirror

 

3.14.设置访问权限

打开/data1/gpdata/master/gpseg-1/pg_hba.conf 按照最下面的格式添加客户端ip或网段

#user define

host    all     all     192.168.1.0/24   trust

host    all     all     127.0.0.1/28    trust

 

3.15.访问方式

可以通过gpAdmin桌面客户端来访问，也可以用命令行来访问，下面来说一下命令行访问的方式，loach是后面添加的用户

[gpadmin@node01 ~]$ psql -d hx_gp -h node01 -p 5432 -U gpadmin

[gpadmin@node01 ~]$ psql -d hx_gp -h node01 -p 5432 -U loach

 

3.16.创建用户

通过命令行登录以后，执行下面的命令

CREATE ROLE loach WITH LOGIN;

ALTER ROLE loach WITH PASSWORD 'loach';

 

 

标签：

版权申明：本站文章部分自网络，如有侵权，请联系：west999com@outlook.com
特别注意：本站所有转载文章言论不代表本站观点，本站所提供的摄影照片，插画，设计作品，如需使用，请与原作者联系，版权归原作者所有