greenplum 集群安装配置（生产环境）

发布时间：2020-05-28 11:20:19 来源：网络阅读：7691 作者：无锋剑栏目：大数据

集群系统初始化信息：

https://blog.51cto.com/michaelkang/2167195

本文对敏感信息进行了替换！！！！

下载软件包：

cd /workspace/gpdb/

wget dl.#kjh#.com/greenplum-db-5.10.2-rhel7-x86_64.rpm

RPM方式，

默认安装到 /usr/local/greenplum-db/ （以root在master执行）
rpm -Uvh greenplum-db-5.10.2-rhel7-x86_64.rpm

建立节点信任关系

打通master到各节点的免密码登录（以root在master执行）

source /usr/local/greenplum-db/greenplum_path.sh

gpssh-exkeys -f /workspace/gpdb/gp-all.txt

安装各个节点，并建立gpadmin用户（以root在master执行）

gpseginstall -f /workspace/gpdb/gp-all.txt -u gpadmin -p gpadmin

安装后检查

source /usr/local/greenplum-db/greenplum_path.sh

gpcheck -f /workspace/gpdb/gp-all.txt -m gpnode615.kjh.com  gpnode616.kjh.com \
-s gpnode611.kjh.com gpnode612.kjh.com  gpnode613.kjh.com gpnode614.kjh.com

创建服务安装文件夹（root用户master节点执行）

#gpssh工具可以同时在多个机器上执行相同命令，和ansible功能类似

/usr/local/gpdata :master 数据数据目录
/data/gpdata :sg 数据数据目录
/data01/gpdata :sg 数据数据目录

gpssh -f /workspace/gpdb/gp-all.txt
=>

groupadd -g 1888 gpadmin
useradd -g 1888 -u 1888 -m -d /home/gpadmin -s /bin/bash gpadmin
chown -R gpadmin.gpadmin /home/gpadmin
echo "gpadmin" | passwd --stdin gpadmin

mkdir -p /usr/local/gpdata/gpmaster
mkdir -p /data/gpdata/primary
mkdir -p /data01/gpdata/primary

mkdir /data/gpdata/mirror
mkdir /data01/gpdata/mirror

chown -R gpadmin:gpadmin /usr/local/gpdata
chown -R gpadmin:gpadmin /data/gpdata
chown -R gpadmin:gpadmin /data01/gpdata

验证

=> df -h|grep data
[gpnode611.kjh.com] /dev/sdb        894G   33M  894G    1% /data01
[gpnode611.kjh.com] /dev/sda        894G   33M  894G    1% /data
[gpnode612.kjh.com] /dev/sda        932G   33M  932G    1% /data01
[gpnode612.kjh.com] /dev/sdb        894G   33M  894G    1% /data
[gpnode614.kjh.com] /dev/sda        932G   33M  932G    1% /data01
[gpnode614.kjh.com] /dev/sdb        932G   33M  932G    1% /data
[gpnode613.kjh.com] /dev/sdb        932G   33M  932G    1% /data
[gpnode613.kjh.com] /dev/sda        932G   33M  932G    1% /data01

设置gpadmin账户的环境（gpadmin账户操作，master节点配置）

修改.bashrc

su - gpadmin 
cat >>/home/gpadmin/.bashrc<<-EOF
source /usr/local/greenplum-db/greenplum_path.sh 
export MASTER_DATA_DIRECTORY=/usr/local/gpdata/gpmaster/gpseg-1 
export PGPORT=5432
EOF

修改.bash_profile

su - gpadmin 
cat >>/home/gpadmin/.bash_profile<<-EOF
source /usr/local/greenplum-db/greenplum_path.sh
export MASTER_DATA_DIRECTORY=/usr/local/gpdata/gpmaster/gpseg-1
export PGPORT=5432
EOF

使之生效

source ~/.bashrc
source ~/.bash_profile

segment 节点 (/etc/hosts文件中的hostname或alias一致)

cat >/usr/local/greenplum-db/all_segment<<-EOF
gpnode611.kjh.com 
gpnode612.kjh.com 
gpnode613.kjh.com 
gpnode614.kjh.com
EOF

验证

more  /usr/local/greenplum-db/all_segment     
gpnode611.kjh.com 
gpnode612.kjh.com 
gpnode613.kjh.com 
gpnode614.kjh.com

初始化数据集群,（gpadmin账户操作，master节点配置）

增加gpinitsystem_config文件为模板

cp $GPHOME/docs/cli_help/gpconfigs/gpinitsystem_config /usr/local/greenplum-db/gpinit_config

配置gpinit_config

vi /usr/local/greenplum-db/gpinit_config

ARRAY_NAME="PT Greenplum Data Platform"
SEG_PREFIX=gpseg
PORT_BASE=40000
declare -a DATA_DIRECTORY=(/data/gpdata/primary /data/gpdata/primary /data01/gpdata/primary /data01/gpdata/primary )
MASTER_HOSTNAME=gpnode615.kjh.com
MASTER_DIRECTORY=/usr/local/gpdata/gpmaster
MASTER_PORT=5432
MIRROR_PORT_BASE=50000
REPLICATION_PORT_BASE=41000
MIRROR_REPLICATION_PORT_BASE=51000
TRUSTED_SHELL=ssh
CHECK_POINT_SEGMENTS=8
ENCODING=UNICODE
DATABASE_NAME=testgpdb
MACHINE_LIST_FILE=/usr/local/greenplum-db/all_segment
declare -a MIRROR_DATA_DIRECTORY=(/data/gpdata/mirror /data/gpdata/mirror /data01/gpdata/mirror /data01/gpdata/mirror)

执行初始化,(master节点 gpadmin 用户执行)

gpinitsystem -c gpinit_config -h hostfile_segments

gpinitsystem命令参数解释：
-c：指定初始化文件。
-h：指定segment主机文件。
-s：指定standby主机，创建standby节点。

su - gpadmin

gpinitsystem -c /usr/local/greenplum-db/gpinit_config -h /usr/local/greenplum-db/all_segment -s gpnode616.kjh.com

初始化过程：

输出集群初始化配置：

=>[INFO]:-Greenplum Database Creation Parameters
=>[INFO]:---------------------------------------
=>[INFO]:-Master Configuration
=>[INFO]:---------------------------------------
=>[INFO]:-Master instance name       = PT Greenplum Data Platform
=>[INFO]:-Master hostname            = gpnode615.kjh.com
=>[INFO]:-Master port                = 5432
=>[INFO]:-Master instance dir        = /usr/local/gpdata/gpmaster/gpseg-1
=>[INFO]:-Master LOCALE              = en_US.utf8
................
=>[INFO]:-Master IP address [4]      = 172.20.6.15
=>[INFO]:-Master IP address [5]      = 172.20.6.221
................
=>[INFO]:-Standby Master             = gpnode616.kjh.com
=>[INFO]:-Primary segment #          = 4
=>[INFO]:-Standby IP address         = ::1
=>[INFO]:-Standby IP address         = 172.20.6.16
=>[INFO]:-Standby IP address         = 172.20.6.220
....................
=>[INFO]:----------------------------------------
=>[INFO]:-Greenplum Primary Segment Configuration
=>[INFO]:----------------------------------------
=>[INFO]:-gpnode611.kjh.com    /data/gpdata/primary/gpseg0   40000   2       0       41000
=>[INFO]:-gpnode611.kjh.com    /data/gpdata/primary/gpseg1   40001   3       1       41001
=>[INFO]:-gpnode611.kjh.com    /data01/gpdata/primary/gpseg2         40002   4       2       41002
.........................
=>[INFO]:-gpnode614.kjh.com    /data01/gpdata/primary/gpseg14        40002   16      14      41002
=>[INFO]:-gpnode614.kjh.com    /data01/gpdata/primary/gpseg15        40003   17      15      41003
=>[INFO]:---------------------------------------
=>[INFO]:-Greenplum Mirror Segment Configuration
=>[INFO]:---------------------------------------
=>[INFO]:-gpnode612.kjh.com    /data/gpdata/mirror/gpseg0    50000   18      0       51000
=>[INFO]:-gpnode612.kjh.com    /data/gpdata/mirror/gpseg1    50001   19      1       51001
.........................
=>[INFO]:-gpnode611.kjh.com    /data01/gpdata/mirror/gpseg14         50002   32      14      51002
=>[INFO]:-gpnode611.kjh.com    /data01/gpdata/mirror/gpseg15         50003   33      15      51003

查看配置正确，确认是否执行？

Continue with Greenplum creation Yy|Nn (default=N):

Y

输出初始化完成概要信息：

=>
=>[INFO]:------------------------------------------------
=>[INFO]:-Parallel process exit status
=>[INFO]:------------------------------------------------
=>[INFO]:-Total processes marked as completed           = 16
=>[INFO]:-Total processes marked as killed              = 0
=>[INFO]:-Total processes marked as failed              = 0
=>[INFO]:------------------------------------------------
=>[INFO]:-Commencing parallel build of mirror segment instances
............................
=>
=>[INFO]:------------------------------------------------
=>[INFO]:-Parallel process exit status
=>[INFO]:------------------------------------------------
=>[INFO]:-Total processes marked as completed           = 16
=>[INFO]:-Total processes marked as killed              = 0
=>[INFO]:-Total processes marked as failed              = 0
=>[INFO]:------------------------------------------------

.....................略
=>
=>]:-Process results...
=>]:-   Successful segment starts                                            = 32
=>]:-   Failed segment starts                                                = 0
=>]:-   Skipped segment starts (segments are marked down in configuration)   = 0
=>]:-----------------------------------------------------
=>]:-Successfully started 32 of 32 segment instances 
=>]:-----------------------------------------------------
.
=>]:------------------------------------------------------
=>]:-Greenplum standby master initialization parameters
=>]:------------------------------------------------------
=>]:-Greenplum master hostname               = gpnode615.kjh.com
=>]:-Greenplum master data directory         = /usr/local/gpdata/gpmaster/gpseg-1
=>]:-Greenplum master port                   = 5432
=>]:-Greenplum standby master hostname       = gpnode616.kjh.com
=>]:-Greenplum standby master port           = 5432
=>]:-Greenplum standby master data directory = /usr/local/gpdata/gpmaster/gpseg-1
=>]:-Greenplum update system catalog         = On

集群初始化成功输出信息如下：

=>:-Greenplum Database instance successfully created
=>]:-------------------------------------------------------
=>]:-To complete the environment configuration, please 
=>]:-update gpadmin .bashrc file with the following
=>]:-1. Ensure that the greenplum_path.sh file is sourced
=>]:-2. Add "export MASTER_DATA_DIRECTORY=/usr/local/gpdata/gpmaster/gpseg-1"
=>]:-   to access the Greenplum scripts for this instance:
=>]:-   or, use -d /usr/local/gpdata/gpmaster/gpseg-1 option for the Greenplum scripts
=>]:-   Example gpstate -d /usr/local/gpdata/gpmaster/gpseg-1
=>]:-Script log file = /home/gpadmin/gpAdminLogs/gpinitsystem_20180903.log
=>]:-To remove instance, run gpdeletesystem utility
=>]:-Standby Master gpnode616.kjh.com has been configured
=>]:-To activate the Standby Master Segment in the event of Master
=>]:-failure review options for gpactivatestandby
=>]:-------------------------------------------------------
=>]:-The Master /usr/local/gpdata/gpmaster/gpseg-1/pg_hba.conf post gpinitsystem
=>]:-has been configured to allow all hosts within this new
=>]:-array to intercommunicate. Any hosts external to this
=>]:-new array must be explicitly added to this file
=>]:-Refer to the Greenplum Admin support guide which is
=>]:-located in the /usr/local/greenplum-db/./docs directory
=>]:-------------------------------------------------------

系统登录初始化设置：

psql修改数据库密码:

psql -d testgpdb
    gpdb=# alter user gpadmin with password 'gpadmin';
    gpdb=# \q

#### 远程登陆

psql -p 5432 -d testgpdb 

# 查询当前版本
select * from version();

设置监听IP和Port

vi ${MASTER_DATA_DIRECTORY}/postgresql.conf

# 设置监听IP (* 生产环境慎用)
listen_addresses = '${ host ip address } '
port = 5432

#### 设置数据库白名单
vi ${MASTER_DATA_DIRECTORY}/pg_hba.conf

# 增加如下一条记录，允许用户密码模式和可以登陆的IP地址范围；
#TYPE  DATABASE  USER   CIDR-ADDRESS     METHOD
host    all     gpadmin     0.0.0.0/0        md5

# 测试登陆

psql -p 5432 -h 172.20.6.15 -U gpadmin -d testgpdb

至此集群配置完成.

可以使用这篇文章的命令查看集群状态：
https://blog.51cto.com/michaelkang/2169857

参考信息：

https://gpdb.docs.pivotal.io/43170/install_guide/init_gpdb.html
https://gp-docs-cn.github.io/docs/utility_guide/admin_utilities/gprecoverseg.html

向AI问一下细节