os: ubunbu 16.04
postgresql: 9.6.8
repmgr: 4.1.1

192.168.56.101 node1
192.168.56.102 node2

Install postgresql 9.6 software

Both node1 and node2 need to use apt install to install postgresql9.6.8. The last bit represents the patch number

# apt install postgresql-plpython-9.6 postgresql-plperl-9.6 postgresql-pltcl-9.6 postgresql-doc-9.6 postgresql-contrib-9.6 postgresql-client-9.6 postgresql-client-common postgresql-common postgresql-9.6 postgresql-server-dev-9.6 libpq-dev libpq5

The advantage of using apt or yum is that the dependency package will be installed, which is more convenient.
After installation, initdb will automatically initialize the database cluster. The main thing is that the files under PG ﹣ data on node2 need to be deleted, because the data needs to be synchronized from node1.

More detailed process.

Set PG? Bindir

Both node1 and node2 need to be set

# vi /etc/profile

PG_BINDIR=/usr/lib/postgresql/9.6/bin

Install repmgr

Both node1 and node2 need to install 2ndQuadrant public apt repository for Debian/Ubuntu

# curl https://dl.2ndquadrant.com/default/release/get/deb | sudo bash
# apt install postgresql-9.6-repmgr repmgr-common

Create user and database

Operation on node1

$ createuser -s repmgr
$ createdb repmgr -O repmgr
$ psql -c "alter user repmgr with password 'repmgrrepmgr'"
$ psql -c "alter user repmgr set search_path to repmgr, \"\$user\", public";

Modify pg_hba.conf

node1 and node2 need to be operated. In fact, the configuration file of each node in the stream replication environment should be the same, which is convenient for switching

$ vi /etc/postgresql/9.6/main/pg_hba.conf
local	repmgr		repmgr					        trust
host	repmgr		repmgr		127.0.0.1/32		trust
host	repmgr		repmgr		192.168.56.0/24		trust

local   replication   repmgr                              trust
host    replication   repmgr      127.0.0.1/32            trust
host    replication   repmgr      192.168.56.0/24         trust

Modify postgresql.conf

node1 and node2 need to be operated. In fact, the configuration file of each node in the stream replication environment should be the same, which is convenient for switching
It needs to be set to the parameter recommended by stream replicator. One of the parameters needs to be modified so as not to restart the instance after modification.

$ vi /etc/postgresql/9.6/main/postgresql.conf
shared_preload_libraries = 'repmgr,pg_stat_statements'

/etc/repmgr.conf

The file contents of node1 node are as follows:

$ vi /etc/repmgr.conf 
node_id=1
node_name=node1
conninfo='host=192.168.56.101 user=repmgr dbname=repmgr connect_timeout=2'
data_directory='/var/lib/postgresql/9.6/main'

The file contents of node2 node are as follows:

$ vi /etc/repmgr.conf 
node_id=2
node_name=node2
conninfo='host=192.168.56.102 user=repmgr dbname=repmgr connect_timeout=2'
data_directory='/var/lib/postgresql/9.6/main'

Register the primary server

Operation on node1 node

$ repmgr -f /etc/repmgr.conf primary register
$ repmgr -f /etc/repmgr.conf cluster show
$ psql -U repmgr
repmgr=# SELECT * FROM repmgr.nodes;

Clone the standby server

Operation on node2 node

$ repmgr -h 192.168.56.101 -U repmgr -d repmgr -f /etc/repmgr.conf standby clone --dry-run
$ repmgr -h 192.168.56.101 -U repmgr -d repmgr -f /etc/repmgr.conf standby clone

– dry run means

--dry-run                           show what would happen for action, but don't execute it

Verify replication is functioning

View on node1 node

$ psql -U repmgr
repmgr=# SELECT * FROM pg_stat_replication;

View on node2 node

$ psql -U repmgr
repmgr=# select * from pg_stat_wal_receiver;

Register the standby

$ repmgr -f /etc/repmgr.conf standby register
$ repmgr -f /etc/repmgr.conf cluster show
 ID | Name  | Role    | Status    | Upstream | Location | Connection string                                              
----+-------+---------+-----------+----------+----------+-----------------------------------------------------------------
 1  | node1 | primary | * running |          | default  | host=192.168.56.101 user=repmgr dbname=repmgr connect_timeout=2
 2  | node2 | standby |   running | node1    | default  | host=192.168.56.102 user=repmgr dbname=repmgr connect_timeout=2
 

You can see that the result of cluster show meets the expectation.

The next blog introduces the switchover of manual repmgr
Reference resources:
https://repmgr.org/docs/4.1/index.html