基于patroni+etcd打造可自动故障转移的PostgreSQL集群

Posted 2021-11-03 PostgreSQLChina

作者：杭州美创科技有限公司

得益于PostgreSQL的开源特性，越来越多的第三方集群管理软件填补了PostgreSQL在集群方面的易用性和可靠性，patroni+etcd提供了一系列的集群管理方案。etcd负责集群状态信息的存放，用来联系各个节点，patroni负责为集群提供高可用服务，两者的集合为PostgreSQL集群提供了故障转移的高可用服务，它不仅配置简单，而且功能丰富：

• 支持手动和自动故障转移
• 支持一主多从、级联复制
• 支持同步、异步模式
• 支持使用watchdog防止脑裂

前期准备

节点规划。实验过程我们使用一主两从构建一套高可用环境。

关闭主机防火墙

# systemctl stop firewalld.service
# systemctl disable firewalld.service

安装postgresql并搭建流复制环境（此步骤略）

在各个节点上部署etcd

安装必要的依赖包及etcd软件

# yum install -y gcc python-devel epel-release
# yum install -y etcd

编辑配置文件（以下列出了需要修改的参数，并以主节点为例）

# vim /etc/etcd/etcd.conf

#[Member]
ETCD_DATA_DIR="/var/lib/etcd/node1.etcd"
ETCD_LISTEN_PEER_URLS="http://192.168.22.128:2380"
ETCD_LISTEN_CLIENT_URLS="http://192.168.22.128:2379,http://127.0.0.1:2379"
#[Clustering]
ETCD_INITIAL_ADVERTISE_PEER_URLS="http://192.168.22.128:2380"
ETCD_ADVERTISE_CLIENT_URLS="http://192.168.22.128:2379"
ETCD_INITIAL_CLUSTER="node1=http://192.168.22.128:2380,node2=http://192.168.22.129:2380, node3=http://192.168.22.130:2380"

启动etcd集群，并设置开机自启动

# systemctl start etcd
# systemctl enable etcd

在各个节点上部署python3

需要使用高版本的python来使用patroni服务，一般的linux环境内置了2.7版本的python环境，因此我们需要升级python，这里采用源码编译安装方式安装

# wget -c https://www.python.org/ftp/python/3.8.2/Python-3.8.2.tar.xz
# ./configure
# make
# make install

删除原2.7版本的软连接，添加新的软链接以使用python3

# rm -f /usr/bin/python
# ln -s /usr/local/bin/python3 /usr/bin/python

在各个节点上部署patroni

安装必要的依赖包和patroni软件

# pip3 install psycopg2-binary -i https://mirrors.aliyun.com/pypi/simple/
# pip3 install patroni -i https://mirrors.aliyun.com/pypi/simple/

修改patroni配置文件（以主节点为例）

# vim /etc/patroni.yml

scope: pgsql
namespace: /pgsql/
name: pgsql_node2
 
restapi:
  listen: 192.168.22.128:8008
  connect_address: 192.168.22.128:8008
 
etcd:
  host: 192.168.22.128:2379
 
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
    postgresql:
      use_pg_rewind: true
      use_slots: true
      parameters:
        listen_addresses: "0.0.0.0"
        port: 5432
        wal_level: logical
        hot_standby: "on"
        wal_keep_segments: 100
        max_wal_senders: 10
        max_replication_slots: 10
        wal_log_hints: "on"
#        archive_mode: "on"
#        archive_timeout: 1800s
#        archive_command: gzip < %p > /data/backup/pgwalarchive/%f.gz
#      recovery_conf:
#        restore_command: gunzip < /data/backup/pgwalarchive/%f.gz > %p
 
postgresql:
  listen: 0.0.0.0:5432
  connect_address: 192.168.22.128:5432
  data_dir: /pgdata/patr2
  bin_dir: /usr/pgsql-12/bin
#  config_dir: /etc/postgresql/9.6/main
  authentication:
    replication:
      username: repl
      password: repl
    superuser:
      username: postgres
      password: postgres
 
#watchdog:
#  mode: automatic # Allowed values: off, automatic, required
#  device: /dev/watchdog
#  safety_margin: 5
 
tags:
    nofailover: false
    noloadbalance: false
    clonefrom: false
    nosync: false

配置patroni服务单元

# vim /etc/systemd/system/patroni.service 
[Unit]
Description=Runners to orchestrate a high-availability PostgreSQL
After=syslog.target network.target

[Service]
Type=simple
User=postgres
Group=postgres
#StandardOutput=syslog
ExecStart=/usr/local/bin/patroni /etc/patroni.yml
ExecReload=/bin/kill -s HUP $MAINPID
KillMode=process
TimeoutSec=30
Restart=no

[Install]
WantedBy=multi-user.target

启动patroni服务

# systemctl start patroni

当然地，我们也可以直接使用patroni命令来启动patroni服务，配置服务单元是为了更方便使用。

#/usr/local/bin/patroni /etc/patroni.yml > patroni.log 2>&1 &

集群环境使用

查看节点信息

# patronictl -c /etc/patroni.yml list
+ Cluster: pgsql (6972099274779350082)+------+---------+----+-----------+
|   Member    |        Host         |  Role  |  State  | TL | Lag in MB |
+-------------+---------------------+--------+---------+----+-----------+
| pgsql_node1 | 192.168.22.128:5432 | Leader | running | 3  |           |
| pgsql_node2 | 192.168.22.129:5432 |        | running | 3  |      0    |
| pgsql_node3 | 192.168.22.130:5432 |        | running | 3  |      0    |
+-------------+---------------------+--------+---------+----+-----------+

手动切换主从，选择某一可用的从节点，使其成为主节点角色

# patronictl -c /etc/patroni.yml switchover
Master [pgsql_node1]: pgsql_node1
Candidate ['pgsql_node2', 'pgsql_node3'] []: pgsql_node2
When should the switchover take place (e.g. 2021-06-20T11:42 )  [now]: now

查看集群状态

# patronictl -c /etc/patroni.yml list
+ Cluster: pgsql (6972099274779350082)+------+---------+----+-----------+
|   Member    |        Host         |  Role  |  State  | TL | Lag in MB |
+-------------+---------------------+--------+---------+----+-----------+
| pgsql_node1 | 192.168.22.128:5432 |        | running | 3  |      0    |
| pgsql_node2 | 192.168.22.129:5432 | Leader | running | 3  |           |
| pgsql_node3 | 192.168.22.130:5432 |        | running | 3  |      0    |
+-------------+---------------------+--------+---------+----+-----------+

自动切换主从。重启node1节点所在主机。查看集群状态，node2自动提升为主，如果只是关闭节点实例，则patroni会再将数据库服务自动拉起。

# patronictl -c /etc/patroni.yml list
+ Cluster: pgsql (6972099274779350082)+------+---------+----+-----------+
|   Member    |        Host         |  Role  |  State  | TL | Lag in MB |
+-------------+---------------------+--------+---------+----+-----------+
| pgsql_node2 | 192.168.22.129:5432 | Leader | running | 3  |           |
| pgsql_node3 | 192.168.22.130:5432 |        | running | 3  |      0    |
+-------------+---------------------+--------+---------+----+-----------+

初始化某一节点。当某一节点与主库不同步，或者节点异常运行时，可以使用此方法初始化节点信息以重新加入集群。

# patronictl -c /etc/patroni.yml reinit pgsql
+ Cluster: pgsql (6972099274779350082)+------+---------+----+-----------+
|   Member    |        Host         |  Role  |  State  | TL | Lag in MB |
+-------------+---------------------+--------+---------+----+-----------+
| pgsql_node1 | 192.168.22.128:5432 |        | running | 3  |      0    |
| pgsql_node2 | 192.168.22.129:5432 | Leader | running | 3  |           |
| pgsql_node3 | 192.168.22.130:5432 |        | running | 3  |      0    |
+-------------+---------------------+--------+---------+----+-----------+
选择以下需要添加的节点名称：pgsql_node3
你确定要重新初始化成员 pgsql_node3？[y/N]：y
成功：为成员pgsql_node3执行初始化

patroni是基于python开发的模板，etcd集群是按照Raft算法和协议开发的，是一个强一致性的、分布式的key-value数据库。两者的结合相辅相成，使得PostgreSQL集群在使用和维护过程中变得更便捷、更透明。