下载PostgreSQL源码包，放在任意目录
设置/etc/sysctl.conf,增加以下内容
kernel.shmmni= 4096 kernel.sem =501000 6412800000 501000 12800 fs.file-max =767246 net.ipv4.ip_local_port_range= 1024 65000 net.core.rmem_default= 1048576 net.core.rmem_max= 1048576 net.core.wmem_default= 262144 net.core.wmem_max= 262144 net.ipv4.tcp_tw_recycle=1 net.ipv4.tcp_max_syn_backlog=4096 net.core.netdev_max_backlog=10000 vm.overcommit_memory=0 net.ipv4.ip_conntrack_max=655360 sysctl -p 生效
如果这里出现最后一个参数未生效可：
modprobeip_conntrack echo"modprobe ip_conntrack" >> /etc/rc.local
设置/etc/security/limits.conf 增加以下内容
* soft nofile 131072 * hard nofile 131072 * soft nproc 131072 * hard nproc 131072 * soft core unlimited * hard core unlimited * soft memlock 50000000 * hard memlock 50000000 --vi/etc/pam.d/login sessionrequired pam_limits.so
设置环境变量
#vi .bash_profile 在文件尾添加 exportPATH=$PATH:/usr/local/pgsql/bin exportMANPATH=$MANPATH:/usr/local/pgsql/man exportLD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/pgsql/lib export PGDATA=/opt/pgdata
先修改好两台机器的名称
[root@primary ~]# groupadd postgres [root@primary ~]# useradd -g postgres postgres [root@primary ~]# passwd postgres [postgres@primary~]$ tar -zxvf postgresql-9.3.0.tar.gz [postgres@primary~]$ cd postgresql-9.3.0 [postgres@primary postgresql-9.3.0]$ ./configure --prefix /home/proxy_pg/pgsql #配置安装目录 [postgres@primary postgresql-9.3.0]$ make #这里需要安装基础的库(gcc、readline、zlib、) 可以不必理会，make时会提示 [root@primary ~]# mkdir /home/proxy_pg [root@primary ~]# chown -R postgres /home/proxy_pg [postgres@primary postgresql-9.3.0]$ make install#安装 [postgres@primary pgsql]$ mkdir /home/proxy_pg/pgsql/data #在你想要存放数据的地方创建data文件夹 [postgres@primary bin]$ ./initdb -D ../data/ #初始化数据库 [postgres@primarydata]$ vi postgresql.conf #修改data/postgresql.conf port = 5432 #可以任意更改你想要的,这里直接用5432 wal_level = hot_standby #主库设置成 hot_standby ，从库才能以READ-ONLY模式打开 max_wal_senders = 30 #WAL STREAM 日志发送进程数 log_connections = on #设置日志参数，记录连接信息 logging_collector = on #开启csvlog输出功能(默认为off) log_directory = 'pg_log' #设置csv日志输出目录($PGDATA目录下，如果不存在该目录，则会自动新建) log_destination = 'csvlog' #设置输出的日志式(类型)，pg中的日志类型有stderr、csvlog、syslog以及eventlog，本例中设置为csvlog #该选项默认为stderr类型 archive_mode= on archive_command = 'test ! -f /archive/pg93/archive_active || cp %p/archive/pg93/%f' #备注：归档标识文件 archive_active 需要手工创建。 wal_keep_segments = 1024 #主库设置的 wal_keep_segments 参数为 1024，一个较大的 wal_keep_segments 设置，允许 备库在宕机较长的时间内依然能够重新追上主库，当然这与主库的繁忙程度有关，主库越忙， 产生的 WAL 日志越多，之前的 WAL 日志越容易被覆盖，进而出现FATAL,XX000错误。 注：修改完logging_collector后需要重启数据库(当然你也可以继续配置，在所有配置都完成后再重启) [root@primary ~]#mkdir -p /archive/pg93 [root@primary ~]# chown -R postgres /archive [root@primary ~]# cd /archive/pg93/ [root@primary pg93]#touch archive_active
以上操作备库也执行
修改data/pg_hba.conf
host replication postgres 192.168.5.0/24 trust
启动主数据库bin/pg_ctl start -D ../data/
[postgres@primarybin]$ ./pg_ctl start -D ../data
检测数据库是否启动成功
[postgres@primarybin]$ ./psql -d postgres psql(9.3.0) Type"help" for help. postgres=# #说明启动成功了
基础备份
基本流程：在主数据库服务器上执行pg_start_backup(),复制data目录，在执行pg_stop_backup()。
./psql–d postgres
postgres=#select pg_start_backup('base backup for logstreaming');
#使用这个方法后，所有请求在写日志之后不会再刷新到磁盘。直到执行pg_stop_backup()这个函数。
#下面需要拷贝一份data目录，并通过scp复制到子数据库中
[postgres@primarypgsql]$ cp -r data data_bak
创建从数据库(standby)
#通过scp方式拷贝data_bac目录到从数据库下(当然也可以通过其他方式)
[postgres@primarypgsql]$ scp -r data_bak/postgres@192.168.5.252:/home/proxy_pg/pgsql/
#进入从数据库服务器，进入刚刚拷贝过来的data_bac目录下
[postgres@standby~]$ cd /home/proxy_pg/pgsql/data_bak
#修改postgres.conf
port = 5432 #改成你想的端口 hot_standby = on #从库上可以执行只读操作 log_connections = on
copy完后，停止主数据库基础备份
postgres=#select pg_stop_backup(); #增加recovery.conf配置下连接的主数据库信息(ip、端口、用户) [postgres@standby share]$ cp recovery.conf.sample ../data/recovery.conf #可以从share中拷贝一份模板 [postgres@standbydata_bak]$ vi recovery.conf standby_mode= 'on' primary_conninfo= 'host=192.168.5.248 port=5432 user=postgres' trigger_file = '/tmp/postgresql.trigger.5432'
#删除pid文件
[postgres@standbydata_bak]$ rm -f postmaster.pid
启动从数据库，并观察CSVLOG
bin/pg_ctl start -D ../data_bak/
观察CSVLOG
[postgres@primarypg_log]$ tail -f postgresql-2014-10-28_155549.csv 2014-10-2815:55:49.522 GMT-8,,,27977,,544f4c05.6d49,1,,2014-10-28 15:55:49GMT-8,,0,LOG,00000,"ending log output to stderr",,"Future logoutput will go to log destination""csvlog"".",,,,,,,"" 2014-10-2815:55:49.523 GMT-8,,,27979,,544f4c05.6d4b,1,,2014-10-28 15:55:49GMT-8,,0,LOG,00000,"database system was shut down at 2014-10-28 15:33:33GMT-8",,,,,,,,,"" 2014-10-2815:55:49.612 GMT-8,,,27977,,544f4c05.6d49,2,,2014-10-28 15:55:49GMT-8,,0,LOG,00000,"database system is ready to acceptconnections",,,,,,,,,"" 2014-10-2815:55:49.612 GMT-8,,,27983,,544f4c05.6d4f,1,,2014-10-28 15:55:49GMT-8,,0,LOG,00000,"autovacuum launcher started",,,,,,,,,"" 2014-10-2815:56:34.436 GMT-8,,,28048,"",544f4c32.6d90,1,"",2014-10-2815:56:34 GMT-8,,0,LOG,00000,"connection received:host=[local]",,,,,,,,,"" 2014-10-2815:56:34.437GMT-8,"postgres","postgres",28048,"[local]",544f4c32.6d90,2,"authentication",2014-10-2815:56:34 GMT-8,2/1,0,LOG,00000,"connection authorized: user=postgresdatabase=postgres",,,,,,,,,"" 2014-10-2816:06:14.597 GMT-8,,,28199,"",544f4e76.6e27,1,"",2014-10-2816:06:14 GMT-8,,0,LOG,00000,"connection received: host=192.168.5.252port=49654",,,,,,,,,"" 2014-10-2816:06:14.598GMT-8,"postgres","",28199,"192.168.5.252:49654",544f4e76.6e27,2,"authentication",2014-10-2816:06:14 GMT-8,3/21,0,LOG,00000,"replication connection authorized:user=postgres",,,,,,,,,"" 2014-10-2816:11:29.501 GMT-8,,,28251,"",544f4fb1.6e5b,1,"",2014-10-2816:11:29 GMT-8,,0,LOG,00000,"connection received: host=192.168.5.252port=50285",,,,,,,,,"" 2014-10-2816:11:29.502GMT-8,"postgres","",28251,"192.168.5.252:50285",544f4fb1.6e5b,2,"authentication",2014-10-2816:11:29 GMT-8,3/28,0,LOG,00000,"replication connection authorized:user=postgres",,,,,,,,,"" 这些信息说明流复制已经成功，从库正准备接收主库的WAL-STREAM。
主库观察WAL-Sender进程
[postgres@primarybin]$ ps -ef |grep post
postgres 2825127977 0 16:11 ? 00:00:00 postgres: wal sender processpostgres 192.168.5.252(50285) streaming 0/60009A8
说明：将输出结果省略部分，可以看到 " wal sender process repuser"进程
在从库上观察 WAL-接收进程
[postgres@standbydata_bak]$ ps -ef |grep post
postgres 1191611911 0 16:11 ? 00:00:00 postgres: wal receiverprocess streaming 0/60009A8
说明：同样省略部分输出结果，可以看到“ wal receiver process ” 进程。
测试
主库上创建用户
postgres=# CREATE ROLE browser LOGIN ENCRYPTED PASSWORD 'browser' postgres-# nosuperuser noinherit nocreatedb nocreateroleCONNECTION LIMIT 200; CREATE ROLE
从库上验证
postgres=# \du
List of roles
Role name | Attributes | Member of
-----------+------------------------------------------------+-----------
browser | No inheritance +| {}
| 200 connections |
postgres | Superuser, Create role, Create DB, Replication | {}
说明：果然，在从库上就立刻创建了新用户 'browser'
主库上创建表空间(On Master)
[postgres@primarydata]$ mkdir -p/home/proxy_pg/pgsql/data/pg_tbs/tbs_browser
从库上也执行 mkdir -p (On Sandby)
[postgres@standbydata_bak]$ mkdir -p /home/proxy_pg/pgsql/data/pg_tbs/tbs_browser
主库上
postgres=# create tablespace tbs_browser owner browser LOCATION postgres-# '/home/proxy_pg/pgsql/data/pg_tbs/tbs_browser'; CREATE TABLESPACE
--在从库上验证
postgres=# \db List of tablespaces Name | Owner | Location -------------+----------+---------------------------------------------- pg_default | postgres | pg_global | postgres | tbs_browser | browser |/home/proxy_pg/pgsql/data/pg_tbs/tbs_browser
(3 rows)
表空间"tbs_browser" 也立刻创建过来了
主库上创建数据库
postgres=# CREATE DATABASE browser WITH OWNER = browser TEMPLATE= template0 ENCODING = 'UTF8' TABLESPACE = tbs_browser; CREATE DATABASE
--从库上验证
postgres=# \l List ofdatabases Name | Owner | Encoding | Collate | Ctype | Access privileges -----------+----------+----------+-------------+-------------+----------------------- browser | browser | UTF8 | zh_CN.UTF-8 | zh_CN.UTF-8 | postgres | postgres | UTF8 |zh_CN.UTF-8 | zh_CN.UTF-8 | template0 | postgres | UTF8 | zh_CN.UTF-8 | zh_CN.UTF-8 |=c/postgres + | | | | | postgres=CTc/postgres template1 | postgres | UTF8 | zh_CN.UTF-8 | zh_CN.UTF-8 |=c/postgres + | | | | | postgres=CTc/postgres (4 rows)
从库上数据库 "browser" 也立刻有了，几乎没有延时。
在从库上建表
postgres=# create table table3(id integer); ERROR: cannot execute CREATE TABLE in a read-onlytransaction
说明：从库是以只读形式打开，只能执行读操作，不能写。
监控streaming
postgres=# select * from pg_stat_replication ; pid |usesysid | usename | application_name| client_addr | client_hostname | client_port | backend_start | state | sent_location | write_location |flush_location | replay_location | sync_priority | sync_state -------+----------+----------+------------------+---------------+-----------------+-------------+-------------------------------+-----------+--------------- +----------------+----------------+-----------------+---------------+------------ 28251 | 10 | postgres | walreceiver |192.168.5.252 | | 50285 | 2014-10-28 16:11:29.501421+08 |streaming | 0/6001C70 | 0/6001C70 | 0/6001C70 | 0/6001C70 | 0 | async (1 row)
本文出自 “岁伏” 博客，请务必保留此出处http://suifu.blog.51cto.com/9167728/1575880
