postgresql集群搭建
基于pgpool中間件實(shí)現(xiàn)postgresql一主多從集群部署,這里用兩臺(tái)服務(wù)器作一主一從示例
虛擬機(jī)名 | IP | 主從劃分 |
---|---|---|
THApps | 192.168.1.31 | 主節(jié)點(diǎn) |
YY-Test-01 | 192.168.1.36 | 從節(jié)點(diǎn) |
vip | 192.168.1.100 | 虛擬ip |
1.軟件版本
系統(tǒng)版本 | CentOS Linux release 7.9.2009 (Core) |
---|---|
PostgreSQL | 10 |
pgpool-II | 4.0.1 |
2.整體架構(gòu)
3.安裝postgresql
3.0.安裝postgresql-10
# 1.更新源
網(wǎng)頁打開:https://yum.postgresql.org/repopackages.php
找到版本 CentOS 7 - x86_64
右鍵 復(fù)制鏈接
# 在centos系統(tǒng)中運(yùn)行:
yum install https://download.postgresql.org/pub/repos/yum/reporpms/EL-7-x86_64/pgdg-redhat-repo-latest.noarch.rpm
# 2.安裝postgresql
# 執(zhí)行命令:
yum update -y
# 查看postgresql源:
yum list | grep postgresql
# 我們需要安裝的是這兩個(gè): postgresql10-contrib postgresql10-server
yum install postgresql10 postgresql10-contrib postgresql10-server postgresql10-devel -y
# 3.初始化數(shù)據(jù)庫 執(zhí)行:
/usr/pgsql-10/bin/postgresql-10-setup initdb
# 4.啟動(dòng)數(shù)據(jù)庫并設(shè)置開機(jī)啟動(dòng)
sudo systemctl start postgresql-10
sudo systemctl enable postgresql-10.service
# 5.安裝完畢,配置pg環(huán)境變量
vi ~/.bash_profile
# 在PATH后添加
/usr/pgsql-10/bin
# 比如:
PATH=$PATH:$HOME/bin:/usr/pgsql-10/bin
# 添加完后執(zhí)行:使它立即生效
source ~/.bash_profile
3.1主節(jié)點(diǎn)配置
(1)創(chuàng)建用于主從訪問的用戶, 修改postgres用戶的密碼,用于遠(yuǎn)程登錄。(切換到postgres用戶操作)
su - postgres
psql
create role actorcloud login replication encrypted password 'public';
alter role postgres with password 'postgres';
(2)修改pg_hba.conf和postgresql.conf配置:
vim /var/lib/pgsql/10/data/pg_hba.conf
在最后一行添加下面內(nèi)容:
host replication actorcloud 192.168.1.31/32 trust
host replication actorcloud 192.168.1.36/32 trust
host all all 192.168.1.0/24 md5
host all all 0.0.0.0/0 md5
修改postgresql.conf配置:
vim /var/lib/pgsql/10/data/postgresql.conf
listen_addresses = '*'
port = 5432
wal_level = replica
max_wal_senders= 10
wal_keep_segments = 10240
max_connections = 512
(3)重啟主節(jié)點(diǎn)
建議先停止,再啟動(dòng),而不是重啟。之后再驗(yàn)證一下是否啟動(dòng)成功
systemctl stop postgresql-10
systemctl start postgresql-10
3.2從節(jié)點(diǎn)配置
從節(jié)點(diǎn)的操作建議全部在postgres用戶下進(jìn)行。
(1)切換postgres用戶
su - postgres
(2)對(duì)主節(jié)點(diǎn)的數(shù)據(jù)進(jìn)行備份到從節(jié)點(diǎn),其中192.168.1.31為主節(jié)點(diǎn)IP,actorcloud是上一節(jié)主節(jié)點(diǎn)創(chuàng)建的用戶。
rm -rf /var/lib/pgsql/10/data/*
pg_basebackup -h 192.168.1.31 -U actorcloud -D /var/lib/pgsql/10/data -X stream -P
(3)拷貝recovery.conf,編輯recovery.conf內(nèi)容,其中192.168.1.31對(duì)應(yīng)主機(jī)IP,actorcloud是上一節(jié)主機(jī)創(chuàng)建的用戶。
cp /usr/pgsql-10/share/recovery.conf.sample /var/lib/pgsql/10/data/recovery.conf
vim /var/lib/pgsql/10/data/recovery.conf
修改配置:
standby_mode = on
primary_conninfo = 'host=192.168.1.31 port=5432 user=actorcloud password=public'
recovery_target_timeline = 'latest'
trigger_file = '/tmp/trigger_file0'
(4)修改從節(jié)點(diǎn)的postgresql.conf,用于開啟standby模式。
hot_standby = on
(5)退出postgres用戶,重啟PostgreSQL
4.驗(yàn)證主從
4.1查看從節(jié)點(diǎn)信息
在主節(jié)點(diǎn)切換至psql界面,輸入命令:
select client_addr,sync_state from pg_stat_replication;
若從節(jié)點(diǎn)顯示出來,如下圖所示,則說明PostgreSQL集群搭建成功。
4.2讀寫測(cè)試
(1)在主節(jié)點(diǎn)寫數(shù)據(jù),從節(jié)點(diǎn)讀數(shù)據(jù)。
create database test;
l
在從節(jié)點(diǎn)上查看創(chuàng)建之后的數(shù)據(jù)庫??梢钥匆姡瑪?shù)據(jù)庫同步了。
l
5.postgres用戶間免密登錄
(0)關(guān)閉防火墻和SETLINUX
# 查看防火墻狀態(tài)
firewall-cmd --state
# 停止firewall
systemctl stop firewalld.service
# 禁止firewall開機(jī)啟動(dòng)
systemctl disable firewalld.service
# 修改/etc/selinux/config 文件
# 將
SELINUX=enforcing
# 改為
SELINUX=disabled
# 重啟
reboot
以下操作兩臺(tái)機(jī)器同時(shí)進(jìn)行:
(1)給postgres用戶更改密碼:密碼修改為postgres
passwd postgres
(2)生成并同步密鑰
以192.168.1.31為例,想要免密登錄另外一臺(tái)機(jī)器。先切換到postgres用戶執(zhí)行,ssh-keygen時(shí)一路回車往下,提升輸入y時(shí)輸入y,ssh-copy-id時(shí)需要輸入密碼。
su postgres
ssh-keygen
ssh-copy-id 192.168.1.36
測(cè)試免密登錄:
ssh 192.168.1.36
ip addr
同理在192.168.1.36機(jī)器上也是相同操作
6.安裝pgpool
(1)安裝(兩臺(tái)機(jī)器都需要安裝)
# 添加源
yum install http://www.pgpool.net/yum/rpms/4.0/redhat/rhel-7-x86_64/pgpool-II-release-4.0-1.noarch.rpm
# 安裝
yum install pgpool-II-pg10-debuginfo pgpool-II-pg10-devel pgpool-II-pg10-extensions pgpool-II-pg10
安裝完之后pgpool的配置文件在/etc/pgpool-II/下
(2)修改配置(兩臺(tái)機(jī)器都需要配置)
pool_hba.conf和之前配置的PostgreSQL中的配置時(shí)一樣的
vim /etc/pgpool-II/pool_hba.conf
host replication actorcloud 192.168.1.31/32 trust
host replication actorcloud 192.168.1.36/32 trust
host all all 192.168.1.0/24 md5
host all all 0.0.0.0/0 md5
(3)對(duì)postgres的密碼進(jìn)行加密。本文將postgres的密碼設(shè)置為和用戶名相同,將加密結(jié)果復(fù)制,并粘貼到pcp.conf中相應(yīng)的位置,取消掉該行的注釋。
pg_md5 postgres
vim /etc/pgpool-II/pcp.conf
postgres:81dc9bdb52d04dc20036dbd8313ed055
(4)執(zhí)行命令(先切換到postgres用戶再執(zhí)行然后輸入密碼):
pg_md5 -m -p -u postgres pool_passwd
7.修改集群配置
以192.168.1.31主節(jié)點(diǎn)為例:
vim /etc/pgpool-II/pgpool.conf
(1)修改監(jiān)聽地址,將localhost改為*,即監(jiān)聽所有地址發(fā)來的請(qǐng)求。
修改前:listen_addresses = 'localhost'
修改后:listen_addresses = '*'
(2)修改backend相關(guān)參數(shù),對(duì)應(yīng)的是PostgreSQ兩個(gè)節(jié)點(diǎn)的相關(guān)信息。
修改前:
backend_hostname0 = 'localhost'
backend_port0 = 5432
backend_weight0 = 1
backend_data_directory0 = '/var/lib/pgsql/data'
backend_flag0 = 'ALLOW_TO_FAILOVER'
修改后:
backend_hostname0 = '192.168.1.31'
backend_port0 = 5432
backend_weight0 = 1
backend_data_directory0 = '/var/lib/pgsql/10/data'
backend_flag0 = 'ALLOW_TO_FAILOVER'
backend_hostname1 = '192.168.1.36'
backend_port1 = 5432
backend_weight1 = 1
backend_data_directory1 = '/var/lib/pgsql/10/data'
backend_flag1 = 'ALLOW_TO_FAILOVER'
(3)pg_hba.conf生效
修改前:enable_pool_hba = off
修改后:enable_pool_hba = on
(4)使負(fù)載均衡生效
修改前:load_balance_mode = off
修改后:load_balance_mode = on
(5)主從流復(fù)制生效,并配置用于檢查的用戶,這個(gè)用戶就用上方創(chuàng)建的用于主從訪問的用戶。
修改前:
master_slave_mode = off
sr_check_period = 0
sr_check_user = 'nobody'
sr_check_password = ''
sr_check_database = 'postgres'
delay_threshold = 0
修改后:
master_slave_mode = on
sr_check_period = 6
sr_check_user = 'actorcloud'
sr_check_password = 'public'
sr_check_database = 'postgres'
delay_threshold = 10000000
(6)健康檢查相關(guān)配置,并配置用于檢查的用戶,這個(gè)用戶就用上方創(chuàng)建的用于主從訪問的用戶。
修改前:
health_check_period = 0
health_check_user = 'nobody'
health_check_password = ''
health_check_database = ''
修改后:
health_check_period = 10
health_check_user = 'actorcloud'
health_check_password = 'public'
health_check_database = 'postgres'
(7)配置主機(jī)故障觸發(fā)執(zhí)行的腳本。
修改前:failover_command = ''
修改后:failover_command = '/var/lib/pgsql/10/failover_stream.sh %d %H'
(8)開啟開門狗,IP為本機(jī)IP
修改前:
use_watchdog = off
wd_hostname = ''
修改后:
use_watchdog = on
wd_hostname = '192.168.1.31'
(9)開啟虛擬IP,并修改網(wǎng)卡信息。啟動(dòng)后直接使用虛擬IP進(jìn)行數(shù)據(jù)庫操作劇哦
修改前:
delegate_IP = ''
if_up_cmd = 'ip addr add $_IP_$/24 dev eth0 label eth0:0'
if_down_cmd = 'ip addr del $_IP_$/24 dev eth0'
arping_cmd = 'arping -U $_IP_$ -w 1 -I eth0'
修改后:
delegate_IP = '192.168.1.100'
if_up_cmd = 'ip addr add $_IP_$/24 dev ens192 label ens192:0'
if_down_cmd = 'ip addr del $_IP_$/24 dev ens192'
arping_cmd = 'arping -U $_IP_$ -w 1 -I ens192'
上方把所有eth0改為了ens192,ens192是該主機(jī)的網(wǎng)卡設(shè)備名稱,可通過命令ip addr
查看
(10)心跳檢查的配置與看門狗配置。IP為從節(jié)點(diǎn)的IP
修改前:
heartbeat_destination0 = 'host0_ip1'
heartbeat_device0 = ''
#other_pgpool_hostname0 = 'host0'
#other_pgpool_port0 = 5432
#other_wd_port0 = 9000
修改后:
heartbeat_destination0 = '192.168.1.36'
heartbeat_device0 = 'ens192'
other_pgpool_hostname0 = '192.168.1.36'
other_pgpool_port0 = 9999
other_wd_port0 = 9000
以上,是主節(jié)點(diǎn)的配置。對(duì)于從節(jié)點(diǎn),只有(8)和(10)的IP需要更改,其他的一致。
8.編寫故障切換腳本
vim /var/lib/pgsql/10/failover_stream.sh
內(nèi)容為:
#! /bin/sh
# Failover command for streaming replication.
# Arguments: $1: new master hostname.
failed_node=$1
new_master=$2
trigger_file=$3
# Do nothing if standby goes down.
if [ $failed_node = 1 ]; then
exit 0;
fi
# Create the trigger file.
# use commond
/usr/bin/ssh -T $new_master /usr/pgsql-10/bin/pg_ctl promote -D /var/lib/pgsql/10/data/
# use file
# /usr/bin/ssh -T $new_master /bin/touch /tmp/trigger_file0
exit 0
9.文件權(quán)限更改
chmod u+s /sbin/ifconfig &&chmod u+s /usr/sbin
chown postgres:postgres /var/lib/pgsql/10/failover_stream.sh && chmod 777 /var/lib/pgsql/10/failover_stream.sh
chown -R postgres.postgres /etc/pgpool-II
mkdir /var/log/pgpool
chown -R postgres.postgres /var/log/pgpool
mkdir /var/run/pgpool
chown -R postgres.postgres /var/run/pgpool
10.運(yùn)行pgpool
# 啟動(dòng)
systemctl start pgpool
# 開機(jī)自啟
systemctl enable pgpool
登錄虛擬ip查看集群節(jié)點(diǎn)。以后所有的操作均可通過連接虛擬IP操作數(shù)據(jù)庫。
psql -p 9999 -h 192.168.1.100 -U postgres
show pool_nodes;
至此,基于 Pgpool-II
中間件的 PostgreSQL
集群搭建完成。
本文摘自 :https://www.cnblogs.com/