fidanf.postgresql_ha

PostgreSQL 高可用性

构建状态

安装和配置一个使用 repmgr 管理的 PostgreSQL 高可用性集群。添加依赖项、扩展、数据库和用户。它也适用于独立安装。

测试环境:

  • Debian 10.x :heavy_check_mark:
  • Debian 11.x :heavy_check_mark:
  • Ubuntu 18.04.x :heavy_check_mark:
  • Ubuntu 20.04.x :heavy_check_mark:

要求

  • Python >=3.8
  • Ansible-core >=2.12

有关用于开发角色的详细依赖项,请参阅 ./requirements.txt

推荐:每个 PostgreSQL 主机

  • Python3 在路径中
  • Pip3 在路径中
  • psycopg2-binary (需要 libpq-dev apt 包)

您可以查看 prepare.yml 来了解 Python 3 的示例设置。

安装

通过命令行

ansible-galaxy collection install community.crypto
ansible-galaxy install fidanf.postgresql_ha

通过 requirements 文件

---
roles:
- src: https://github.com/fidanf/ansible-role-postgresql-ha
  name: fidanf.postgresql_ha
  version: master

collections:
  - community.crypto

依赖项

示例清单

有关详细的组和主机变量,请查看 inventory.yml

角色变量

角色的默认变量分为两个文件:

为了确切了解这些变量的用途和有效值,请随时检查 templates 目录中的所有 Jinja 模板。原始的默认配置文件也已包含(.orig)。

示例剧本

---
- hosts: pgcluster
  gather_facts: yes
  become: yes
  roles:
    - name: fidanf.postgresql_ha
      vars:
        # 必需的配置项
        repmgr_target_group: pgcluster
        repmgr_master: pgsql01
        repmgr_failover: automatic
        repmgr_promote_command: /usr/bin/repmgr standby promote -f /etc/repmgr.conf --log-to-file
        repmgr_follow_command: /usr/bin/repmgr standby follow -f /etc/repmgr.conf --log-to-file --upstream-node-id=%n
        repmgr_monitoring_history: "yes"
        repmgr_connection_check_type: query
        repmgr_log_level: DEBUG
        repmgr_reconnect_attempts: 2
        repmgr_reconnect_interval: 10
        # 基本设置
        postgresql_version: 13
        postgresql_cluster_name: main
        postgresql_cluster_reset: false # TODO: 需要对 repmgr 测试
        postgresql_listen_addresses: "*"
        postgresql_port: 5432
        postgresql_wal_level: "replica"
        postgresql_max_wal_senders: 10
        postgresql_max_replication_slots: 10
        postgresql_wal_keep_segments: 100
        postgresql_hot_standby: on
        postgresql_archive_mode: on
        postgresql_archive_command: "test ! -f /tmp/%f && cp %p /tmp/%f"
        postgresql_ext_install_repmgr: yes
        postgresql_shared_preload_libraries:
          - repmgr
        # PostgreSQL 日志
        postgresql_log_checkpoints: on
        postgresql_log_connections: on
        postgresql_log_disconnections: on
        postgresql_log_temp_files: 0
        # pg_hba.conf
        postgresql_pg_hba_custom:
          - { type: "host", database: "all", user: "all", address: "192.168.56.0/24", method: "md5" }
          - { type: "host", database: "replication", user: "{{ repmgr_user }}", address: "192.168.56.0/24", method: "trust" }  
          - { type: "host", database: "replication", user: "{{ repmgr_user }}", address: "127.0.0.1/32", method: "trust" }  
          - { type: "host", database: "{{ repmgr_database }}", user: "{{ repmgr_user }}", address: "127.0.0.1/32", method: "trust" }  
          - { type: "host", database: "{{ repmgr_database }}", user: "{{ repmgr_user }}", address: "192.168.56.0/32", method: "trust" }  
        # 数据库
        postgresql_databases:
          - name: "{{ repmgr_database }}"
            owner: "{{ repmgr_user }}"
            encoding: "UTF-8"
          - name: testdb
            owner: admin
            encoding: "UTF-8"
        # 用户
        postgresql_users:
          - name: "{{ repmgr_user }}"
            pass: "{{ repmgr_password }}"
          - name: admin
            pass: secret # PostgreSQL >=10 不接受未加密密码
            encrypted: yes
        # 角色
        postgresql_user_privileges:
          - name: "{{ repmgr_user }}"
            db: "{{ repmgr_database }}"
            priv: "ALL"
            role_attr_flags: "SUPERUSER,REPLICATION"
          - name: admin
            db: testdb
            role_attr_flags: "SUPERUSER"

首次安装后可运行的有用命令

验证集群功能

ansible pgcluster -b --become-user postgres -m shell -a "repmgr cluster crosscheck"

显示集群状态

ansible pgcluster -b --become-user postgres -m shell -a "repmgr cluster show"

列出节点及其属性

ansible pgcluster -b --become-user postgres -m shell -a "repmgr node status"

注册(克隆)额外的备用节点

# 假设当前主节点的主机名为 pgsql01
ansible-playbook myplaybook.yml -l 'pgsql04' -e 'repmgr_primary_hostname=pgsql01' -vv 

在自动故障转移后将前主节点注册为备用节点

postgres@pgsql01:~$ pg_ctlcluster 13 main stop
postgres@pgsql01:~$ repmgr standby clone --force -h pgsql02 -U repmgr -d repmgr
postgres@pgsql01:~$ pg_ctlcluster 13 main start
postgres@pgsql01:~$ repmgr standby register --force

或者您可以使用 repmgr node rejoin 和 pg_rewind

repmgr node rejoin -d repmgr -U repmgr -h pgsql02 --verbose --force-rewind=/usr/lib/postgresql/13/bin/pg_rewind

许可证

MIT / BSD

关于项目

Install and configure PostgreSQL cluster managed with repmgr. Add dependencies, extensions, databases and users.

安装
ansible-galaxy install fidanf.postgresql_ha
许可证
mit
下载
676
拥有者
DevOps Engineer ~ Ansible lover & CI/CD builder. I'm learning to more effectively deploy and monitor apps every day.