如何在不停机的情况下向现有的 PostgreSQL 级联复制添加额外的从属服务器？

步骤 1：将第 3 个从属信息放在从属 2 的 pg_hba.conf 中，如果这里没有 recovery.conf 文件，请立即创建

cat recovery.conf
standby_mode='on'
primary_conninfo = 'host=192.168.2.3 port=5432 user=postgres password=postgres'

#restore_command = 'cp /archive/%f "%p"'
#archive_cleanup_command = '/usr/pgsql-9.4/bin/pg_archivecleanup /archive %r'
recovery_target_timeline = 'latest'

Pg_hba.conf:

host    replication     postgres            192.168.2.5/32  trust
host    replication     postgres            192.168.2.4/32  trust

步骤2. reload或者重新启动slave_2服务器：

Option 1: From the command-line shell
su - postgres
/usr/bin/pg_ctl -D /DATA reload

Option 2: Using SQL
SELECT pg_reload_conf();

步骤3：暂停slave_2服务器，然后检查是否暂停

select pg_wal_replay_pause();
select pg_is_wal_replay_paused();

注意：如果您的服务器正在测试，则在暂停slave_2后，只需在主服务器上创建一些表，因为恢复后您需要检查slave_2和slave_3是否复制了数据 

步骤4：在slave_2上进行物理备份并将其发送到slave_3：

ON SLAVE_3:
/usr/pgsql-10/bin/pg_ctl -D /var/lib/pgsql/10/data/ stop  -----stop the slave_3 cluster
rm -vfr /var/lib/pgsql/10/data/*      ---remove the old data's

FROM SLAVE_2:
rsync -cva --inplace /DATA/* [email protected]:/var/lib/pgsql/10/data/

slave_3 pg_hba.conf:
host    replication     postgres            192.168.2.4/3 2  trust

Slave_3 postgresql.conf:
listen_addresses = 'localhost,192.168.2.5'
port = 5432 
wal_level = 'hot_standby'       # minimal, archive, hot_standby, or logical
max_wal_senders = 10             # max number of walsender processes
                                # (change requires restart)
wal_keep_segments = 64          # in logfile segments, 16MB each; 0 disables
hot_standby = on
archive_mode = on               # allows archiving to be done
archive_command =  'rsync -av %p /archive/%f && rsync -av %p [email protected]:/archive/%f' # or 'cd .'

slave_3 recovery.conf:
cat /var/lib/pgsql/10/data/recovery.conf
standby_mode = 'on'      # to enable the standby (read-only) mode.
primary_conninfo = 'host=192.168.2.4 port=5432 user=postgres password=postgres'
                         # to specify a connection info to the master node.

步骤5：恢复slave_2服务器然后重新启动slave_3服务器：

At slave_2:
select pg_wal_replay_resume();
At Slave_3:
/usr/pgsql-10/bin/pg_ctl -D /var/lib/pgsql/10/data/ start

监控 SLAVE_3 和 MASTER 服务器：

ON SLAVE_3:
postgres=# select * from pg_stat_wal_receiver ;
  pid  |  status   | receive_start_lsn | receive_start_tli | received_lsn | received_tli |       last_msg_send_time       |      last_msg_receipt_time       | latest_end_lsn |         latest_end_time          | slot_name |                                                                                             conninfo                                                                                             
-------+-----------+-------------------+-------------------+--------------+--------------+--------------------------------+----------------------------------+----------------+----------------------------------+-----------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 39970 | streaming | 0/1A000000        |                 1 | 0/1A000000   |            1 | 2018-03-15 16:39:03.4787+05:30 | 2018-03-15 16:39:01.426584+05:30 | 0/1A30B500     | 2018-03-15 16:37:03.161287+05:30 |           | user=postgres password=******** dbname=replication host=192.168.2.4 port=5432 fallback_application_name=walreceiver sslmode=prefer sslcompression=1 krbsrvname=postgres target_session_attrs=any
(1 row)

ON MASTER:
select * from pg_stat_replication ;
  pid  | usesysid | usename  | application_name | client_addr | client_hostname | client_port |          backend_start           | backend_xmin |   state   |  sent_lsn  | write_lsn  | flush_lsn  | replay_lsn | write_lag | flush_lag | replay_lag | sync_priority | sync_state 
-------+----------+----------+------------------+-------------+-----------------+-------------+----------------------------------+--------------+-----------+------------+------------+------------+------------+-----------+-----------+------------+---------------+------------
 14640 |       10 | postgres | walreceiver      | 192.168.2.3 |                 |       57716 | 2018-03-15 09:33:48.617764+05:30 |              | streaming | 0/1A30B500 | 0/1A30B500 | 0/1A30B500 | 0/1A30B500 |           |           |            |             0 | async
(1 row)