PostgreSQL数据库单机扩展为流复制

primary:10.189.102.118

standby:10.189.100.195

1. 配置ssh互信机制

• 在primary主库执行

$ ssh-keygen -t rsa
$ cp ~/.ssh/id_rsa.pub ~/.ssh/authorized_keys
$ ssh-copy-id postgres@10.189.100.195

• 在standby备库执行

$ ssh-keygen -t rsa
$ cp ~/.ssh/id_rsa.pub ~/.ssh/authorized_keys
$ ssh-copy-id postgres@10.189.102.118

2. 在standby服务器安装postgres数据库,不需要初始化.

安装过程详见:http://www.cnblogs.com/ilifeilong/p/6979288.html

3. 在primary服务器创建具有REPLICATION权限的复制用户

postgres=# CREATE ROLE repl WITH REPLICATION PASSWORD 'repl' LOGIN;

4. 允许复制用户远程连接到primary服务器

$ grep "^host" pg_hba.conf
host    all             all             127.0.0.1/32            trust
host    replication             repl             0.0.0.0/0               md5 
host    all             all             ::1/128                 trust

5. 在primary服务器设置流复制相关的参数

$ mkdir /usr/local/pgsql/arch 
$ egrep "archive_mode|max_wal_senders|wal_keep_segments|archive_command|wal_level|hot_standby" postgresql.conf
al_level = hot_standby            # minimal, archive, hot_standby, or logical
archive_mode = on        # enables archiving; off, on, or always
archive_command = 'test ! -f /usr/local/pgsql/arch/%f && cp %p /usr/local/pgsql/arch/%f && rsync -a %p postgres@10.189.100.195:/usr/local/pgsql/arch/%f'        
max_wal_senders = 5        # max number of walsender processes
wal_keep_segments = 30        # in logfile segments, 16MB each; 0 disables
hot_standby = on            # "on" allows queries during recovery
#hot_standby_feedback = off        # send info from standby to prevent

6. 重新启动primary服务器进程

$ pg_ctl stop -m fast
$ pg_ctl start

7. 对primary服务器做一个全备并传输到standby服务器

•    方法一,在primary服务器通过pg_(start|stop)_backup函数进行备份

postgres=# SELECT pg_start_backup('label', true);
 pg_start_backup 
-----------------
 7/E6000060
(1 row)
$ rsync -az --progress ${PGDATA} postgres@10.189.100.195:/usr/local/pgsql/ --exclude postmaster.pid
postgres=# SELECT pg_stop_backup();
NOTICE:  pg_stop_backup complete, all required WAL segments have been archived
 pg_stop_backup 
----------------
 7/E60005C8
(1 row)

•  方法二,在standby服务器通过pg_basebackup命令进行备份,要求standby的PGDATA目录为空

$ pg_basebackup --host=10.189.102.118 --username=repl --port=5432 --label=backup --verbose --progress --pgdata=/usr/local/pgsql/data --checkpoint=fast --format=p --xlog-method=stream
Password: 
transaction log start point: 7/EA000028 on timeline 1
pg_basebackup: starting background WAL receiver
65933562/65933562 kB (100%), 1/1 tablespace                                         
transaction log end point: 7/EA000830
pg_basebackup: waiting for background process to finish streaming ...
pg_basebackup: base backup completed

8. 设置standby数据库复制相关参数,使得standby失效转移后可以作为主库工作

$ mkdir /usr/local/pgsql/arch
$ grep "hot_standby" postgresql.conf
hot_standby = on                        # "on" allows queries during recovery
#hot_standby_feedback = off             # send info from standby to prevent

9. 在standby文件创建恢复文件

$ cat recovery.conf 
restore_command = 'cp /usr/local/pgsql/arch/%f "%p"'
standby_mode = 'on'
primary_conninfo = 'user=repl password=repl host=10.189.102.118 port=5432 sslmode=disable sslcompression=1'
archive_cleanup_command = 'pg_archivecleanup -d /usr/local/pgsql/arch %r >> /usr/local/pgsql/arch/archive_cleanup.log'
trigger_file = '/usr/local/pgsql/data/trigger_active_standby'

10. 启动standby数据库进程,自动启动流复制

$ pg_ctl start -w
waiting for server to start....LOG:  could not create IPv6 socket: Address family not supported by protocol
LOG:  redirecting log output to logging collector process
HINT:  Future log output will appear in directory "pg_log".
 done
server started

11. 检查primary和standby数据库的状态

• 通过函数和系统表查看

edbstore=# select * from pg_stat_replication;           #在primary主库查看
-[ RECORD 1 ]----+------------------------------
pid              | 15013
usesysid         | 19206
usename          | repl
application_name | walreceiver
client_addr      | 10.189.100.195
client_hostname  | 
client_port      | 56072
backend_start    | 2017-06-13 08:10:35.400508-07
backend_xmin     | 
state            | streaming
sent_location    | 7/EC01A588
write_location   | 7/EC01A588
flush_location   | 7/EC01A588
replay_location  | 7/EC01A588
sync_priority    | 0
sync_state       | async

edbstore=# SELECT pg_current_xlog_location();                      #在primary主库查看
 pg_current_xlog_location 
--------------------------
 7/EC01A588
(1 row)

postgres=# select pg_last_xlog_receive_location(),pg_last_xlog_replay_location(),pg_last_xact_replay_timestamp();     #在standby备库查看
 pg_last_xlog_receive_location | pg_last_xlog_replay_location | pg_last_xact_replay_timestamp 
-------------------------------+------------------------------+-------------------------------
 7/EC01A588                    | 7/EC01A588                   | 2017-06-13 08:25:20.281568-07
(1 row)

• 通过进程查看

$ ps -ef | grep sender | grep -v grep        #在primary库查看
postgres 15013 24883  0 08:10 ?        00:00:00 postgres: wal sender process repl 10.189.100.195(56072) streaming 7/EC01A668

$  ps -ef | grep receiver | grep -v grep     #在standby库查看
postgres 12857 12843  0 08:10 ?        00:00:00 postgres: wal receiver process   streaming 7/EC01A668

•  查看备库落后主库多少字节的WAl日志

postgres=# select pg_xlog_location_diff(pg_current_xlog_location(),replay_location) from pg_stat_replication ;     #在primary库查看
 pg_xlog_location_diff 
-----------------------
                     0
(1 row)

•  把主库WAL日志位置转换成WAL文件名称和偏移量

postgres=# select pg_current_xlog_location();
 pg_current_xlog_location 
--------------------------
 7/EC021790
(1 row)

postgres=# select * from pg_xlogfile_name_offset('7/EC021790');
        file_name         | file_offset 
--------------------------+-------------
 000000010000000700000076 |      137104
(1 row)

• 查看主库备库的状态

postgres=# select pg_is_in_recovery();            #在primary主库查看
 pg_is_in_recovery 
-------------------
 f
(1 row)

postgres=# select pg_is_in_recovery();            #在standby库查看
 pg_is_in_recovery 
-------------------
 t
(1 row)