一、同步流复制的架构
PostgreSQL从9.1版本之后提供了同步复制的架构。同步复制要求在数据写入Standby数据库后，事务的commit才返回,所以Standby库出现问题时，会导致主库被hang住。解决这个问题的方法是启动两个Standby数据库，这两个standby数据库只要有一个是正常的，就不会让主库hang住。所以在实际应用中，同步流复制，总是有1个主库和2个以上的Standby库。
二、同步复制的配置
实现同步复制功能主要是在主库上配置参数"synchronous_standby_names",这个参数指定多个Standby的名称,各个名称通过逗号分隔，而Standby名称是在Standby连接到主库时，由连接参数"application_name"指定的。要使用同步复制，在Standby数据库中,recovery.conf
里的primary_conninfo一定要指定连接参数"application_name"。recovery.conf的示例如下:
standby_mode = ‘on‘
primary_conninfo = ‘application_name=standby01 user=repluesr password=XXXXXX host=172.16.0.111 port=5432 sslmode=disable sslcompression=1‘
三、系统环境
系统平台：centos6.5
PostgreSQL版本：postgresql9.4.3
防火墙关闭/iptables
selinux设置成
SELINUX=disabled
|
主机名 |
IP地址 |
角色 |
数据目录 |
|
postgresql01 |
172.16.0.111 |
主库 |
/home/postgres/pgsql/data |
|
postgresql02 |
172.16.0.112 |
Standby |
/home/postgres/pgsql/data |
|
postgresql03 |
172.16.0.113 |
Standby |
/home/postgres/pgsql/data |
四、源码包安装
1、在三台安装依赖包
yum -y install gcc*
yum -y install readline-devel
2、在三台增加用户
# adduser postgres
# passwd postgres
Changing password for user postgres.
New password:
BAD PASSWORD: it is too simplistic/systematic
BAD PASSWORD: is too simple
Retype new password:
passwd: all authentication tokens updated successfully.
注意：主库安装与配置好后，在把全部的文件拷备到standby中。
以下是主库操作
3. 下载PostgreSQL 源码包
# wget http://ftp.postgresql.org/pub/source/v9.4.3/postgresql-9.4.3.tar.bz2
4. 解压源码包
# tar xjf postgresql-9.4.3.tar.bz2
5. 进入解压后的目录
# cd postgresql-9.4.3
6.开始编译安装PostgreSQL 数据库。
[[email protected] postgresql-9.4.3]# ./configure --prefix=/home/postgres/pgsql
[[email protected] postgresql-9.4.3]#gmake
[[email protected] postgresql-9.4.3]# gmake install
7.设置环境
[[email protected] postgresql-9.4.3]# cd /home/postgres/
[[email protected] postgres]# ls
pgsql
[[email protected] postgres]# vi .bash_profile
把 PATH=$PATH:$HOME/bin
改成 PATH=$PATH:$HOME/bin:/home/postgres/pgsql/bin
保存退出。
让环境变量生效
[[email protected] postgres]# source .bash_profile
8.初始化数据库
8.1新建数据目录
[[email protected] postgres]# mkdir /home/postgres/pgsql/data
8.2更改权限
[[email protected] postgres]# chown postgres:postgres /home/postgres/pgsql/data
8.3切换到postgres用户
[[email protected] postgres]# su - postgres
8.4init db
[[email protected] ~]$ /home/postgres/pgsql/bin/initdb -D /home/postgres/pgsql/data
到这里数据的初始化就完成
9.系统服务
9.1回到root用户
[[email protected] ~]$ exit
9.2复制安装目录下的linux文件/etc/init.d/
进入postgresql的安装目录
[[email protected] postgres]# cd /root/postgresql-9.4.3/
[[email protected] postgresql-9.4.3]# cp contrib/start-scripts/linux /etc/init.d/postgresql
9.3修改/etc/init.d/postgresql 注意：红色是修改部分
[[email protected] postgresql-9.4.3]# vi /etc/init.d/postgresql
#! /bin/sh
# chkconfig: 2345 98 02
# description: PostgreSQL RDBMS
# This is an example of a start/stop script for SysV-style init, such
# as is used on Linux systems. You should edit some of the variables
# and maybe the ‘echo‘ commands.
#
# Place this file at /etc/init.d/postgresql (or
# /etc/rc.d/init.d/postgresql) and make symlinks to
# /etc/rc.d/rc0.d/K02postgresql
# /etc/rc.d/rc1.d/K02postgresql
# /etc/rc.d/rc2.d/K02postgresql
# /etc/rc.d/rc3.d/S98postgresql
# /etc/rc.d/rc4.d/S98postgresql
# /etc/rc.d/rc5.d/S98postgresql
# Or, if you have chkconfig, simply:
# chkconfig --add postgresql
#
# Proper init scripts on Linux systems normally require setting lock
# and pid files under /var/run as well as reacting to network
# settings, so you should treat this with care.
# Original author: Ryan Kirkpatrick <[email protected]>
# contrib/start-scripts/linux
## EDIT FROM HERE
# Installation prefix
prefix=/home/postgres /pgsql
# Data directory
PGDATA="/home/postgres/pgsql/data"
# Who to run the postmaster as, usually "postgres". (NOT "root")
PGUSER=postgres
# Where to keep a log file
PGLOG="$PGDATA/serverlog"
# It‘s often a good idea to protect the postmaster from being killed by the
# OOM killer (which will tend to preferentially kill the postmaster because
# of the way it accounts for shared memory). Setting the OOM_SCORE_ADJ value
# to -1000 will disable OOM kill altogether. If you enable this, you probably
# want to compile PostgreSQL with "-DLINUX_OOM_SCORE_ADJ=0", so that
# individual backends can still be killed by the OOM killer.
#OOM_SCORE_ADJ=-1000
# Older Linux kernels may not have /proc/self/oom_score_adj, but instead
# /proc/self/oom_adj, which works similarly except the disable value is -17.
# For such a system, enable this and compile with "-DLINUX_OOM_ADJ=0".
#OOM_ADJ=-17
## STOP EDITING HERE
# The path that is to be used for the script
PATH=/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin
# What to use to start up the postmaster. (If you want the script to wait
# until the server has started, you could use "pg_ctl start -w" here.
# But without -w, pg_ctl adds no value.)
DAEMON="$prefix/bin/postmaster"
# What to use to shut down the postmaster
PGCTL="$prefix/bin/pg_ctl"
set -e
# Only start if we can find the postmaster.
test -x $DAEMON ||
{
echo "$DAEMON not found"
if [ "$1" = "stop" ]
then exit 0
else exit 5
fi
}
# Parse command line parameters.
case $1 in
start)
echo -n "Starting PostgreSQL: "
test x"$OOM_SCORE_ADJ" != x && echo "$OOM_SCORE_ADJ" > /proc/self/oom_score_adj
test x"$OOM_ADJ" != x && echo "$OOM_ADJ" > /proc/self/oom_adj
su - $PGUSER -c "$DAEMON -D ‘$PGDATA‘ &" >>$PGLOG 2>&1
echo "ok"
;;
stop)
echo -n "Stopping PostgreSQL: "
su - $PGUSER -c "$PGCTL stop -D ‘$PGDATA‘ -s -m fast"
echo "ok"
;;
restart)
echo -n "Restarting PostgreSQL: "
su - $PGUSER -c "$PGCTL stop -D ‘$PGDATA‘ -s -m fast -w"
test x"$OOM_SCORE_ADJ" != x && echo "$OOM_SCORE_ADJ" > /proc/self/oom_score_adj
test x"$OOM_ADJ" != x && echo "$OOM_ADJ" > /proc/self/oom_adj
su - $PGUSER -c "$DAEMON -D ‘$PGDATA‘ &" >>$PGLOG 2>&1
echo "ok"
;;
reload)
echo -n "Reload PostgreSQL: "
su - $PGUSER -c "$PGCTL reload -D ‘$PGDATA‘ -s"
echo "ok"
;;
status)
su - $PGUSER -c "$PGCTL status -D ‘$PGDATA‘"
;;
*)
# Print help
echo "Usage: $0 {start|stop|restart|reload|status}" 1>&2
exit 1
;;
esac
exit 0
9.4启动数据库
[[email protected] postgresql-9.4.3]# /etc/init.d/postgresql start
9.5让数据库开机启动
[[email protected] postgresql-9.4.3]# chkconfig --add postgresql
[[email protected] postgresql-9.4.3]# chkconfig postgresql on
9.6创建数据操作历史记录文件
[[email protected] postgresql-9.4.3]# touch /home/postgres/pgsql/.pgsql_history
[[email protected] postgresql-9.4.3]# chown postgres:postgres /home/postgres/pgsql/.pgsql_history
10.测试使用
[[email protected] ~]$ createdb test
[[email protected] ~]$ psql test
psql (9.4.3)
Type "help" for help.
test=#
源码编译安装成功
五、配置同步流复制
配置如下：
1.在主库增加同步的用户名与密码
[[email protected] ~]$ psql -d postgres
psql (9.4.3)
Type "help" for help.
postgres=# CREATE ROLE repluser REPLICATION LOGIN PASSWORD ‘zhf#2015‘;
CREATE ROLE
postgres=#
2.在主库postgresql01上进行配置
修改以下参数
在主库的/home/postgres/pgsql/data/pg_hba.conf中做如下配置：
[[email protected] ~]$ vi /home/postgres/pgsql/data/pg_hba.conf
增加以下：
host replication repluser 172.16.0.0/24 md5
2.在主库postgreso1的/home/postgres/pgsql/data/postgresql.conf中设置如下三个参数
[[email protected] ~]$ vi /home/postgres/pgsql/data/postgresql.conf
isten_addresses = ‘*‘
max_wal_senders = 5
wal_level = hot_standby
3.在主数据上指定同步复制的Standby名称，在postgresql01的/home/postgres/pgsql/data/postgresql.conf增加如下一行：
synchronous_standby_names = ‘standby01,standby02‘
4.停止主库postgresql
[[email protected] ~]# /etc/init.d/postgresql stop
5.把主库安装文件拷备到两台standby中
[[email protected] ~]# scp -r /home/postgres/pgsql 172.16.0.112:/home/postgres/
[[email protected] ~]# scp -r /home/postgres/pgsql 172.16.0.113:/home/postgres/
[[email protected] ~]# scp -r /etc/init.d/postgresql 172.16.0.112:/etc/init.d/
[[email protected] ~]# scp -r /etc/init.d/postgresql 172.16.0.113:/etc/init.d/
6.在两台standby中设置权限两台都操作如下：
[[email protected] ~]# cd /home/postgres/pgsql/
[[email protected] pgsql]# chown -R postgres:postgres data
[[email protected] pgsql]# chmod +x /etc/init.d/postgresql
[[email protected] postgres]# chkconfig --add postgresql
[[email protected] postgres]# chkconfig postgresql on
[[email protected] ~]# cd /home/postgres/pgsql/
[[email protected] pgsql]# chown -R postgres:postgres data
[[email protected] pgsql]# chmod +x /etc/init.d/postgresql
[[email protected] postgres]# chkconfig --add postgresql
[[email protected] postgres]# chkconfig postgresql on
7. 在两台standby中设置环境变量两台都操作如下：
[[email protected] pgsql]# cd /home/postgres/
把 PATH=$PATH:$HOME/bin
改成 PATH=$PATH:$HOME/bin:/home/postgres/pgsql/bin
保存退出。
让环境变量生效：
[[email protected] postgres]# source .bash_profile
[[email protected] pgsql]# cd /home/postgres/
把 PATH=$PATH:$HOME/bin
改成 PATH=$PATH:$HOME/bin:/home/postgres/pgsql/bin
保存退出。
让环境变量生效：
[[email protected] postgres]# source .bash_profile
8.在备库postgresql02上进行配置
把配置文件拷过来
[[email protected] postgres]# su - postgres
[[email protected] ~]$ cd pgsql/data/
[[email protected] data]$ cp /home/postgres/pgsql/share/recovery.conf.sample ./recovery.conf
在postgresql02的/home/postgres/pgsql/data/recovery.conf中的配置“primary_conninfo”里增加 连接参数“appliction_name”,如下：
standby_mode = on
primary_conninfo = ‘application_name=standby01 user=repluser password=zhf#2015 host=172.16.0.111 port=5432 sslmode=disable sslcompression=1‘
在postgresql02的/home/postgres/pgsql/data/postgresql.conf中的配置参数
hot_standby = on
完成配置后，启动数据库，如下：
[[email protected] postgres]# /etc/init.d/postgresql start
9.在备库postgresql03上进行配置
把配置文件拷过来
[[email protected] postgres]# su - postgres
[[email protected] ~]$ cd pgsql/data/
[[email protected] data]$ cp /home/postgres/pgsql/share/recovery.conf.sample ./recovery.conf
在postgresql03的/home/postgres/pgsql/data/recovery.conf中的配置“primary_conninfo”里增加 连接参数“appliction_name”,如下：
standby_mode = on
primary_conninfo = ‘application_name=standby02 user=repluser password=zhf#2015 host=172.16.0.111 port=5432 sslmode=disable sslcompression=1‘
在postgresql03的/home/postgres/pgsql/data/postgresql.conf中的配置参数
hot_standby = on
完成配置后，启动数据库，如下：
[[email protected] postgres]# /etc/init.d/postgresql start
10.启动主库 postgresql01
11.测试在主库中
[[email protected] data]$ psql test
psql (9.4.3)
Type "help" for help.
test=# create table test01(id int primary key,note text);
CREATE TABLE
test=# insert into test01 values(1,‘1111111‘);
INSERT 0 1
test=# insert into test01 values(2,‘2222222‘);
INSERT 0 1
在备库查询数据
在备库postgresql02中查看
[[email protected] ~]$ psql test
psql (9.4.3)
Type "help" for help.
test=# select * from test01;
id | note
----+---------
1 | 1111111
2 | 2222222
(2 rows)
test=#
在备库postgresql03中查看
[[email protected] postgres]# su - postgres
[[email protected] ~]$ psql test
psql (9.4.3)
Type "help" for help.
test=# select * from test01;
id | note
----+---------
1 | 1111111
2 | 2222222
(2 rows)
test=#