PostgreSQL_配置repmgr(高可用)

目录

主节点环境准备：

主节点配置repmgr

备节点环境准备：

配置主备节点ssh互信

配置repmgrd实现自动故障转移

配置见证节点 防止脑裂（存在多个主节点）

---

主节点环境准备：

 关闭防火墙 防止备节点连接失败

[postgres@localhost conf]$ su root
[root@localhost conf]# firewall-cmd --add-port=5666/tcp --permanent
[root@localhost conf]# firewall-cmd --reload

修改主机名

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

192.168.144.140 myserver01
192.168.144.141 myserver02
192.168.144.145 myserver03
192.168.144.149 myserver04

上传源码包并解压

[postgres@localhost ~]$ cd resoruce/
[postgres@localhost resoruce]$ ll
-rw-r--r--.  1 postgres postgres    453529 May  7 03:18 repmgr-5.3.1.tar.gz
[postgres@localhost resoruce]$ tar -zxvf repmgr-5.3.1.tar.gz 

环境检查

[postgres@localhost resoruce]$ cd repmgr-5.3.1
[postgres@localhost repmgr-5.3.1]$ ./configure

安装依赖

[postgres@localhost repmgr-5.3.1]$ su root
[root@localhost repmgr-5.3.1]# yum install -y flex

编译安装

[postgres@localhost repmgr-5.3.1]$ make
[postgres@localhost repmgr-5.3.1]$ make install

安装完成

[postgres@localhost repmgr-5.3.1]$ ls /home/postgres/soft/bin | grep repmgr
repmgr
repmgrd

创建扩展

[postgres@localhost repmgr-5.3.1]$ psql -U pgadmin -d appdb
appdb=# create extension repmgr;

相关表

appdb=# select * from repmgr.nodes limit 10;
appdb=# select * from repmgr.events limit 10;
appdb=# select * from repmgr.monitoring_history limit 10;
appdb=# select * from repmgr.replication_status limit 10;
appdb=# select * from repmgr.show_nodes limit 10;

主节点配置repmgr

开启复制模式

[postgres@localhost data]$ psql -U pgadmin -d appdb
psql (12.0)
Type "help" for help.

appdb=# alter system set archive_mode = on;
ALTER SYSTEM
appdb=# alter system set archive_command = 'cp %p /home/postgres/data/%f';
ALTER SYSTEM
appdb=# alter system set wal_level = replica;
ALTER SYSTEM
appdb=# alter system set listen_addresses = '*';

 配置pg_hba文件 保证备节点可以连接

[postgres@localhost data]$ vi $PGDATA/pg_hba.conf

host    appdb          postgres    0.0.0.0/0    trust
host    replication    all         0.0.0.0/0    trust

 重启数据库

[postgres@localhost data]$ pg_ctl restart

修改repmgr配置文件

[postgres@localhost data]$ mkdir -p /home/postgres/conf
[postgres@localhost data]$ cd /home/postgres/conf/
[postgres@localhost conf]$ vi repmgr.conf

node_id = 1
node_name = 'myserver01'
conninfo = 'host=myserver01 user=pgadmin dbname=appdb port=5666 password=1qaz@WSX'
data_directory = '/home/postgres/soft/../data'    #这个是备节点准备克隆的目录 与$PGDATA保持一致

log_level = 'DEBUG'
log_facility = 'STDERR'
log_file = '/home/postgres/conf/repmgr.log'
log_status_interval = 150

注册主节点

[postgres@localhost conf]$ repmgr -f /home/postgres/conf/repmgr.conf primary register
INFO: connecting to primary database...
DEBUG: connecting to: "user=pgadmin password=1qaz@WSX dbname=appdb host=myserver01 port=5666 connect_timeout=2 fallback_application_name=repmgr options=-csearch_path="
INFO: "repmgr" extension is already installed
NOTICE: primary node record (ID: 1) registered

查看主节点注册信息

[postgres@localhost conf]$ repmgr cluster show -f /home/postgres/conf/repmgr.conf

备节点环境准备：

 关闭防火墙

[postgres@localhost conf]$ su root
[root@localhost conf]# firewall-cmd --add-port=5666/tcp --permanent
[root@localhost conf]# firewall-cmd --reload

修改主机名

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

192.168.144.140 myserver01
192.168.144.141 myserver02
192.168.144.145 myserver03
192.168.144.149 myserver04

创建用户 

[root@localhost ~]# groupadd postgres
[root@localhost ~]# useradd -g postgres postgres
[root@localhost ~]# passwd postgres

创建表空间目录

[postgres@localhost ~]$ su - root
[root@localhost ~]# cd /
[root@localhost /]# mkdir /exam
[root@localhost /]# chown postgres.postgres /exam

 创建日志目录

[postgres@localhost openssl]$ su root
[root@localhost openssl]# mkdir /pglog
[root@localhost openssl]# chown postgres.postgres /pglog

安装数据库（不用初始化数据库）

[root@localhost ~]# yum install -y openssl-devel gcc gcc-c++ readline-devel zlib-devel

[postgres@myserver02 resource]$ tar -xvf postgresql-12.0.tar
[postgres@myserver02 resource]$ cd postgresql-12.0
[postgres@myserver02 postgresql-12.0]$ ./configure --prefix=/home/postgres/soft/ --with-openssl --with-pgport=5666
[postgres@myserver02 postgresql-12.0]$ make world -j4
[postgres@myserver02 postgresql-12.0]$ make install-world -j4

[postgres@myserver02 postgresql-12.0]$ su - postgres
[postgres@myserver02 postgresql-12.0]$ vi ~/.bashrc
[postgres@myserver02 postgresql-12.0]$ source ~/.bashrc

export PGHOME=/home/postgres/soft
export PGDATA=${PGHOME}/../data
export PATH=${PGHOME}/bin:${PATH}

 安装repmgr

[postgres@myserver02 resource]$ cd ~/resource/
[postgres@myserver02 resource]$ su root
[root@myserver02 resource]# yum install -y flex
[root@myserver02 resource]# su - postgres
[postgres@myserver02 resource]$ tar -zxvf repmgr-5.3.1.tar.gz
[postgres@myserver02 repmgr-5.3.1]$ cd resource/repmgr-5.3.1
[postgres@myserver02 repmgr-5.3.1]$ ./configure
[postgres@myserver02 repmgr-5.3.1]$ make
[postgres@myserver02 repmgr-5.3.1]$ make install

配置repmgr

[postgres@localhost repmgr-5.3.1]$ mkdir -p /home/postgres/conf
[postgres@localhost repmgr-5.3.1]$ cd /home/postgres/conf/
[postgres@localhost conf]$ vi repmgr.conf

node_id = 2
node_name = 'myserver02'
conninfo = 'host=myserver02 user=pgadmin dbname=appdb port=5666 password=1qaz@WSX'
data_directory = '/home/postgres/data'   #这个目录一定要是空目录才行 要往这个目录克隆

log_level = 'DEBUG'
log_facility = 'STDERR'
log_file = '/home/postgres/conf/repmgr.log'
log_status_interval = 150

配置pgpass 不然克隆时会报密码错误

[postgres@localhost soft]$ cd ~
[postgres@localhost ~]$ touch ~/.pgpass
[postgres@localhost ~]$ chmod 0600 ~/.pgpass 
[postgres@localhost ~]$ vi ~/.pgpass

#hostname:port:database:username:password
localhost:5666:postgres:pgadmin:1qaz@WSX

将主节点克隆到备节点(备节点执行)

[postgres@localhost conf]$ repmgr -h 192.168.144.140 -U pgadmin -d appdb -f /home/postgres/conf/repmgr.conf  standby clone --dry-run
[postgres@localhost conf]$ repmgr -h 192.168.144.140 -U pgadmin -d appdb -f /home/postgres/conf/repmgr.conf  standby clone

配置openssl     因为主节点开启了openssl 备节点不开启 会启动失败 

过程见 PostgreSQL开启openssl(通讯加密)_GNAIXGNAHZ的博客-CSDN博客

[postgres@localhost ~]$ cd ~
[postgres@localhost ~]$ mkdir openssl
[postgres@localhost ~]$ cd openssl/
[postgres@localhost openssl]$ openssl req -new -nodes -text -out root.csr -keyout root.key -subj "/CN=test1"
[postgres@localhost openssl]$ chmod og-rwx root.key
[postgres@localhost openssl]$ cp /etc/pki/tls/openssl.cnf /home/postgres/openssl/
[postgres@localhost openssl]$ vi openssl.cnf 

[ CA_default ]
dir             = /home/postgres/openssl                # Where everything is kept
certs           = $dir          # Where the issued certs are kept

[postgres@localhost openssl]$ openssl x509 -req -in root.csr -text -days 3650 -extfile /home/postgres/openssl//openssl.cnf -extensions v3_ca -signkey root.key -out root.crt
[postgres@localhost openssl]$ openssl req -new -nodes -text -out server.csr -keyout server.key -subj "/CN=test1"
[postgres@localhost openssl]$ chmod og-rwx server.key
[postgres@localhost openssl]$ openssl x509 -req -in server.csr -text -days 365 -CA root.crt -CAkey root.key -CAcreateserial -out server.crt
[postgres@localhost openssl]$ vi $PGDATA/postgresql.conf

ssl = on
#ssl_ca_file = ''
ssl_cert_file = '/home/postgres/openssl/server.crt'
#ssl_crl_file = ''
ssl_key_file = '/home/postgres/openssl/server.key'

 修改配置文件 保证备库可以连接

[postgres@localhost conf]$ vi $PGDATA/pg_hba.conf 
host    all             all             192.168.144.0/24           trust
host    replication     all             192.168.144.0/24           trust

[postgres@localhost conf]$ vi $PGDATA/postgresql.conf 
listen_addresses = '*'

启动备节点

[root@localhost openssl]# su postgres
[postgres@localhost openssl]$ pg_ctl start

 注册备节点

[postgres@localhost openssl]$ repmgr -f /home/postgres/conf/repmgr.conf standby register

配置主备节点ssh互信

主节点信任备节点

[postgres@myserver01 conf]$ su - postgres
[postgres@myserver01 ~]$ ssh-keygen
[postgres@myserver01 ~]$ ssh-copy-id postgres@myserver02   #备节点IP

备节点信任主节点

[postgres@myserver02 conf]$ su - postgres
[postgres@myserver02 ~]$ ssh-keygen
[postgres@myserver02 ~]$ ssh-copy-id postgres@myserver01        #主节点IP

将备节点手工切换为主节点（手动故障转移）

[postgres@localhost ~]$ repmgr -f /home/postgres/conf/repmgr.conf standby switchover --siblings-follow
[postgres@localhost ~]$ repmgr cluster -f /home/postgres/conf/repmgr.conf show

配置repmgrd实现自动故障转移

停止集群 先停止备库 后停止主库

[postgres@localhost conf]$ pg_ctl stop

所有节点都需要添加预加载库

[postgres@myserver01 conf]$ vi $PGDATA/postgresql.conf

shared_preload_libraries = 'repmgr'

所有节点修改repmgr配置文件 添加参数

[postgres@localhost conf]$ vi /home/postgres/conf/repmgr.conf

monitor_interval_secs = 1
connection_check_type = ping
reconnect_attempts = 2
reconnect_interval = 5

failover = automatic        
promote_command = '/home/postgres/soft/bin/repmgrd standby promote -f /home/postgres/conf/repmgr.conf'
follow_command  = '/home/postgres/soft/bin/repmgrd standby follow -f /home/postgres/conf/repmgr.conf --upstream-node-id=%n'

启动集群 先启动主节点 后启动备节点

[postgres@localhost conf]$ pg_ctl start

所有节点启动repmgrd

[postgres@localhost conf]$ repmgrd -f /home/postgres/conf/repmgr.conf -d -p /home/postgres/conf/repmgr.pid

检查repmgrd是否正常

[postgres@myserver01 conf]$ repmgr -f /home/postgres/conf/repmgr.conf service status

停止主节点 模拟故障

[postgres@localhost conf]$ pg_ctl stop

配置见证节点 防止脑裂（存在多个主节点）

明天再写

配置系统服务 保证能够使用systemctl接管

[postgres@myserver01 system]$ su root
[root@myserver01 system]# vi /usr/lib/systemd/system/postgresql-12.service

[Unit]
Description=PostgreSQL database server
After=network.target
[Service]
Type=forking
User=postgres
Group=postgres
OOMScoreAdjust=-1000
Environment=PG_OOM_ADJUST_FILE=/proc/self/oom_score_adj
Environment=PG_OOM_ADJUST_VALUE=0
Environment=PGDATA=/home/postgres/data
Environment=PGLOG=/tmp/logfile
ExecStart=/home/postgres/soft/bin/pg_ctl start -D ${PGDATA} -l ${PGLOG}
ExecReload=/bin/kill -HUP $MAINPID
KillMode=mixed
KillSignal=SIGINT
TimeoutSec=0
[Install]
WantedBy=multi-user.target

配置sudoers

[root@myserver01 system]# visudo
[root@myserver01 system]# scp /etc/sudoers root@server2:/etc/  #复制到从节点

#includedir /etc/sudoers.d
Defaults:postgres !requiretty
postgres ALL=NOPASSWD:/bin/systemctl start postgresql-12,\
/bin/systemctl stop postgresql-12,\
/bin/systemctl reload postgresql-12,\
/bin/systemctl status postgresql-12,\
/bin/systemctl restart postgresql-12

 检查sudo是否正常

[root@myserver01 ~]# su postgres
[postgres@myserver01 root]$ sudo systemctl status postgresql-12
[postgres@myserver01 root]$ sudo systemctl stop postgresql-12
[postgres@myserver01 root]$ sudo systemctl start postgresql-12

配置repmgr.conf其他参数

[postgres@myserver01 root]$ vi /home/postgres/conf/repmgr.conf 

#[options configuration]
replication_user = 'appuser'
replication_type = 'physical'
ssh_options='-q -o ConnectTimeout=10'
#[log record]
log_level = 'DEBUG'
log_facility = 'STDERR'
log_file = '/home/postgres/conf/repmgr.log'
log_status_interval = 150
#[service command configuration]
service_start_command   = 'sudo systemctl start postgresql-12'
service_stop_command    = 'sudo systemctl stop postgresql-12'
service_restart_command = 'sudo systemctl restart postgresql-12'
service_reload_command  = 'sudo systemctl reload postgresql-12'