操作系統：CentOS Linux release 7.6.1810 (Core)
node1:192.168.216.130 master
node2:192.168.216.132 slave
node3:192.168.216.136 haproxy
這里僅測試，所以只部署了一主一叢，適用與測試環境，生產環境建議postgres至少1主2從，3個etcd節點,2個haproxy+keepalive組成
一、首先在兩個節點上安裝postgres，下面以postgres9.5.19為例
1、添加RPM yum install https://download.postgresql.org/pub/repos/yum/9.5/redhat/rhel-7-x86_64/pgdg-centos95-9.5-3.noarch.rpm 2、安裝PostgreSQL 9.5 yum install postgresql95-server postgresql95-contrib 注意：本次實驗我們這里只需要操作到第2步即可，初始化可以由patroni來替我們完成 3、初始化數據庫 /usr/pgsql-9.5/bin/postgresql95-setup initdb 4、設置開機自啟動 systemctl enable postgresql-9.5.service 5、啟動服務 systemctl start postgresql-9.5.service 6、查看版本 psql --version
二、安裝etcd服務
1、這里我只在node1單節點上安裝，僅實驗,未做分布式部署，如果集群部署可以參考博客etcd集群部署文章
yum install etcd -y cp /etc/etcd/etcd.conf /etc/etcd/etcd.conf.bak cd /etc/etcd/ [root@localhost etcd]# egrep ^[A-Z] ./etcd.conf ETCD_DATA_DIR="/var/lib/etcd/node1.etcd" ETCD_LISTEN_PEER_URLS="http://192.168.216.130:2380" ETCD_LISTEN_CLIENT_URLS="http://192.168.216.130:2379,http://127.0.0.1:2379" ETCD_NAME="node1" ETCD_INITIAL_ADVERTISE_PEER_URLS="http://192.168.216.130:2380" ETCD_ADVERTISE_CLIENT_URLS="http://192.168.216.130:2379" ETCD_INITIAL_CLUSTER="node1=http://192.168.216.130:2380" ETCD_INITIAL_CLUSTER_TOKEN="etcd-cluster" ETCD_INITIAL_CLUSTER_STATE="new"
2、保存文件，然后重啟etcd服務
systemctl restart etcd
3、查看ectd服務是否正常
三、安裝patroni,分別在node1和node2節點安裝
1、安裝patroni用到依賴包，這里通過pip安裝patroni
yum install gcc yum install python-devel.x86_64 curl https://bootstrap.pypa.io/get-pip.py -o get-pip.py python get-pip.py pip install psycopg2-binary pip install --upgrade setuptools pip install patroni[etcd,consul]
2、驗證patroni是否安裝成功
3、配置patroni，以下操作在node1中進行
mkdir /data/patroni/conf -p cd /data/patroni/conf yum install git git clone https://github.com/zalando/patroni.git cd /data/patroni/conf/patroni-master cp -r postgres0.yml ../conf/
4、編輯node1上的postgres0.yml文件
scope: batman
#namespace: /service/
name: postgresql0
restapi:
listen: 192.168.216.130:8008
connect_address: 192.168.216.130:8008
# certfile: /etc/ssl/certs/ssl-cert-snakeoil.pem
# keyfile: /etc/ssl/private/ssl-cert-snakeoil.key
# authentication:
# username: username
# password: password
# ctl:
# insecure: false # Allow connections to SSL sites without certs
# certfile: /etc/ssl/certs/ssl-cert-snakeoil.pem
# cacert: /etc/ssl/certs/ssl-cacert-snakeoil.pem
etcd:
host: 192.168.216.130:2379
bootstrap:
# this section will be written into Etcd:/<namespace>/<scope>/config after initializing new cluster
# and all other cluster members will use it as a `global configuration`
dcs:
ttl: 30
loop_wait: 10
retry_timeout: 10
maximum_lag_on_failover: 1048576
# master_start_timeout: 300
synchronous_mode: false
#standby_cluster:
#host: 127.0.0.1
#port: 1111
#primary_slot_name: patroni
postgresql:
use_pg_rewind: true
use_slots: true
parameters:
wal_level: logical
hot_standby: "on"
wal_keep_segments: 1000
max_wal_senders: 10
max_replication_slots: 10
wal_log_hints: "on"
archive_mode: "on"
archive_timeout: 1800s
archive_command: mkdir -p ../wal_archive && test ! -f ../wal_archive/%f && cp %p ../wal_archive/%f
recovery_conf:
restore_command: cp ../wal_archive/%f %p
# some desired options for 'initdb'
initdb: # Note: It needs to be a list (some options need values, others are switches)
- encoding: UTF8
- data-checksums
pg_hba: # Add following lines to pg_hba.conf after running 'initdb'
# For kerberos gss based connectivity (discard @.*$)
#- host replication replicator 127.0.0.1/32 gss include_realm=0
#- host all all 0.0.0.0/0 gss include_realm=0
- host replication replicator 0.0.0.0/0 md5
- host all admin 0.0.0.0/0 md5
- host all all 0.0.0.0/0 md5
# Additional script to be launched after initial cluster creation (will be passed the connection URL as parameter)
# post_init: /usr/local/bin/setup_cluster.sh
# Some additional users users which needs to be created after initializing new cluster
users:
admin:
password: postgres
options:
- createrole
- createdb
replicator:
password: replicator
options:
- replication
postgresql:
listen: 0.0.0.0:5432
connect_address: 192.168.216.130:5432
data_dir: /data/postgres
bin_dir: /usr/pgsql-9.5/bin/
# config_dir:
# pgpass: /tmp/pgpass0
authentication:
replication:
username: replicator
password: replicator
superuser:
username: admin
password: postgres
# rewind: # Has no effect on postgres 10 and lower
# username: rewind_user
# password: rewind_password
# Server side kerberos spn
# krbsrvname: postgres
parameters:
# Fully qualified kerberos ticket file for the running user
# same as KRB5CCNAME used by the GSS
# krb_server_keyfile: /var/spool/keytabs/postgres
unix_socket_directories: '.'
#watchdog:
# mode: automatic # Allowed values: off, automatic, required
# device: /dev/watchdog
# safety_margin: 5
tags:
nofailover: false
noloadbalance: false
clonefrom: false
nosync: false
5、配置patroni，以下操作在node2中進行
mkdir /data/patroni/conf -p cd /data/patroni/conf yum install git git clone https://github.com/zalando/patroni.git cd /data/patroni/conf/patroni-master cp -r postgres1.yml ../conf/
6、編輯node2上的postgres1.yml文件
scope: batman
#namespace: /service/
name: postgresql1
restapi:
listen: 192.168.216.132:8008
connect_address: 192.168.216.132:8008
# certfile: /etc/ssl/certs/ssl-cert-snakeoil.pem
# keyfile: /etc/ssl/private/ssl-cert-snakeoil.key
# authentication:
# username: username
# password: password
# ctl:
# insecure: false # Allow connections to SSL sites without certs
# certfile: /etc/ssl/certs/ssl-cert-snakeoil.pem
# cacert: /etc/ssl/certs/ssl-cacert-snakeoil.pem
etcd:
host: 192.168.216.130:2379
bootstrap:
# this section will be written into Etcd:/<namespace>/<scope>/config after initializing new cluster
# and all other cluster members will use it as a `global configuration`
dcs:
ttl: 30
loop_wait: 10
retry_timeout: 10
maximum_lag_on_failover: 1048576
# master_start_timeout: 300
synchronous_mode: false
#standby_cluster:
#host: 127.0.0.1
#port: 1111
#primary_slot_name: patroni
postgresql:
use_pg_rewind: true
use_slots: true
parameters:
wal_level: logical
hot_standby: "on"
wal_keep_segments: 1000
max_wal_senders: 10
max_replication_slots: 10
wal_log_hints: "on"
archive_mode: "on"
archive_timeout: 1800s
archive_command: mkdir -p ../wal_archive && test ! -f ../wal_archive/%f && cp %p ../wal_archive/%f
recovery_conf:
restore_command: cp ../wal_archive/%f %p
# some desired options for 'initdb'
initdb: # Note: It needs to be a list (some options need values, others are switches)
- encoding: UTF8
- data-checksums
pg_hba: # Add following lines to pg_hba.conf after running 'initdb'
# For kerberos gss based connectivity (discard @.*$)
#- host replication replicator 127.0.0.1/32 gss include_realm=0
#- host all all 0.0.0.0/0 gss include_realm=0
- host replication replicator 0.0.0.0/0 md5
- host all admin 0.0.0.0/0 md5
- host all all 0.0.0.0/0 md5
# Additional script to be launched after initial cluster creation (will be passed the connection URL as parameter)
# post_init: /usr/local/bin/setup_cluster.sh
# Some additional users users which needs to be created after initializing new cluster
users:
admin:
password: postgres
options:
- createrole
- createdb
replicator:
password: replicator
options:
- replication
postgresql:
listen: 0.0.0.0:5432
connect_address: 192.168.216.132:5432
data_dir: /data/postgres
bin_dir: /usr/pgsql-9.5/bin/
# config_dir:
# pgpass: /tmp/pgpass0
authentication:
replication:
username: replicator
password: replicator
superuser:
username: admin
password: postgres
# rewind: # Has no effect on postgres 10 and lower
# username: rewind_user
# password: rewind_password
# Server side kerberos spn
# krbsrvname: postgres
parameters:
# Fully qualified kerberos ticket file for the running user
# same as KRB5CCNAME used by the GSS
# krb_server_keyfile: /var/spool/keytabs/postgres
unix_socket_directories: '.'
#watchdog:
# mode: automatic # Allowed values: off, automatic, required
# device: /dev/watchdog
# safety_margin: 5
tags:
nofailover: false
noloadbalance: false
clonefrom: false
nosync: false
7、記下data_dir上述yml配置文件中的值。該目錄需要確保postgres用戶具備寫入的權限。如果此目錄不存在，請創建它：在node1和node2節點分別進行如下操作
mkdir /data/postgres -p chown -Rf postgres:postgres /data/postgres chmod 700 /data/postgres
8、在node1上切換到postgres用戶,並啟動patroni服務,這里patroni會幫我們自動初始化數據庫並創建相應的角色
chown -Rf postgres:postgres /data/patroni/conf su - postgres 啟動patroni服務 patroni /data/patroni/conf/postgres0.yml
此時如果服務正常啟動可以打印以下日志信息
由於不是后台啟動的服務，所以這里我們克隆一個窗口，切換到postgres用戶下，並執行psql -h 127.0.0.1 -U admin postgres連接數據庫，驗證patroni是否正常托管postgres服務
9、在node2上切換到postgres用戶,並啟動patroni服務,這里和node1的操作一致
chown -Rf postgres:postgres /data/patroni/conf su - postgres 啟動patroni服務 patroni /data/patroni/conf/postgres1.yml
如果服務啟動正常，可看到如下日志打印信息
10、查詢集群運行狀態patronictl -c /data/patroni/conf/postgres0.yml list
11、patronictl -c /data/patroni/conf/postgres0.yml switchover 手動切換master
12、可以后台啟動來保持patroni服務不中斷,也可以配置成systemd服務來管理保證開機自啟
node1節點：
nohup patroni /data/patroni/conf/postgres0.yml > /data/patroni/patroni_log 2>&1 &
node2節點：
nohup patroni /data/patroni/conf/postgres1.yml > /data/patroni/patroni_log 2>&1 &
四、在node3節點安裝haproxy
yum install -y haproxy cp -r /etc/haproxy/haproxy.cfg /etc/haproxy/haproxy.cfg_bak
編輯haproxy.cfg配置文件
# vi /etc/haproxy/haproxy.cfg
#---------------------------------------------------------------------
# 全局定義
global
# log語法：log [max_level_1]
# 全局的日志配置，使用log關鍵字，指定使用127.0.0.1上的syslog服務中的local0日志設備，
# 記錄日志等級為info的日志
# log 127.0.0.1 local0 info
log 127.0.0.1 local1 notice
chroot /var/lib/haproxy
pidfile /var/run/haproxy.pid
# 定義每個haproxy進程的最大連接數 ，由於每個連接包括一個客戶端和一個服務器端，
# 所以單個進程的TCP會話最大數目將是該值的兩倍。
maxconn 4096
# 用戶，組
user haproxy
group haproxy
# 以守護進程的方式運行
daemon
# turn on stats unix socket
stats socket /var/lib/haproxy/stats
#---------------------------------------------------------------------
# 默認部分的定義
defaults
# mode語法：mode {http|tcp|health} 。http是七層模式，tcp是四層模式，health是健康檢測，返回OK
mode tcp
# 使用127.0.0.1上的syslog服務的local3設備記錄錯誤信息
log 127.0.0.1 local3 err
#if you set mode to http,then you nust change tcplog into httplog
option tcplog
# 啟用該項，日志中將不會記錄空連接。所謂空連接就是在上游的負載均衡器或者監控系統為了
#探測該服務是否存活可用時，需要定期的連接或者獲取某一固定的組件或頁面，或者探測掃描
#端口是否在監聽或開放等動作被稱為空連接；官方文檔中標注，如果該服務上游沒有其他的負
#載均衡器的話，建議不要使用該參數，因為互聯網上的惡意掃描或其他動作就不會被記錄下來
option dontlognull
# 定義連接后端服務器的失敗重連次數，連接失敗次數超過此值后將會將對應后端服務器標記為不可用
retries 3
# 當使用了cookie時，haproxy將會將其請求的后端服務器的serverID插入到cookie中，以保證
#會話的SESSION持久性；而此時，如果后端的服務器宕掉了，但是客戶端的cookie是不會刷新的
#，如果設置此參數，將會將客戶的請求強制定向到另外一個后端server上，以保證服務的正常
option redispatch
#等待最大時長 When a server's maxconn is reached, connections are left pending in a queue which may be server-specific or global to the backend.
timeout queue 1m
# 設置成功連接到一台服務器的最長等待時間，默認單位是毫秒
timeout connect 10s
# 客戶端非活動狀態的超時時長 The inactivity timeout applies when the client is expected to acknowledge or send data.
timeout client 1m
# Set the maximum inactivity time on the server side.The inactivity timeout applies when the server is expected to acknowledge or send data.
timeout server 1m
timeout check 5s
maxconn 5120
#---------------------------------------------------------------------
# 配置haproxy web監控，查看統計信息
listen status
bind 0.0.0.0:1080
mode http
log global
stats enable
# stats是haproxy的一個統計頁面的套接字，該參數設置統計頁面的刷新間隔為30s
stats refresh 30s
stats uri /haproxy-stats
# 設置統計頁面認證時的提示內容
stats realm Private lands
# 設置統計頁面認證的用戶和密碼，如果要設置多個，另起一行寫入即可
stats auth admin:passw0rd
# 隱藏統計頁面上的haproxy版本信息
# stats hide-version
#---------------------------------------------------------------------
listen master
bind *:5000
mode tcp
option tcplog
balance roundrobin
option httpchk OPTIONS /master
http-check expect status 200
default-server inter 3s fall 3 rise 2 on-marked-down shutdown-sessions
server node1 192.168.216.130:5432 maxconn 1000 check port 8008 inter 5000 rise 2 fall 2
server node2 192.168.216.132:5432 maxconn 1000 check port 8008 inter 5000 rise 2 fall 2
listen replicas
bind *:5001
mode tcp
option tcplog
balance roundrobin
option httpchk OPTIONS /replica
http-check expect status 200
default-server inter 3s fall 3 rise 2 on-marked-down shutdown-sessions
server node1 192.168.216.130:5432 maxconn 1000 check port 8008 inter 5000 rise 2 fall 2
server node2 192.168.216.132:5432 maxconn 1000 check port 8008 inter 5000 rise 2 fall 2
啟動haproxy服務
systemctl start haproxy systemctl status haproxy
瀏覽器訪問http://192.168.216.136:1080/haproxy-stats輸入用戶名admin密碼passw0rd
這里我們通過5000端口和5001端口分別來提供寫服務和讀服務,如果需要對數據庫寫入數據只需要對外提供192.168.216.136:5000即可,可以模擬主庫故障，即關閉其中的master節點來驗證是否會進行自動主從切換
https://www.linode.com/docs/databases/postgresql/create-a-highly-available-postgresql-cluster-using-patroni-and-haproxy/#configure-etcd
https://www.opsdash.com/blog/postgres-getting-started-patroni.html