How to Build a High‑Availability GreatSQL MGR Cluster with Docker‑Compose

For internet‑facing high‑performance, high‑concurrency, high‑availability systems, distributed architecture is an inevitable evolution.

SpringCloud and Service Mesh address application‑layer scaling and deployment.

Redis Cluster and Codis solve distributed caching.

Kubernetes advances the operating‑system layer for better resource efficiency.

In the database layer, sharding, high‑availability, and other governance solutions are moving to distributed databases (NewSQL) as an unstoppable trend.

What Is GreatSQL

GreatSQL is a native distributed relational database featuring dynamic scaling, strong data consistency, and high‑availability clusters. It uses a shared‑nothing architecture with data redundancy, replica management, sharding, and MPP technology to deliver high performance and supports dynamic node expansion. It is widely used in finance, telecom, energy, government, and internet core systems and fully compatible with domestic operating systems, chips, and other hardware.

GreatSQL can serve as a free, drop‑in replacement for MySQL or Percona Server in production environments.

GreatSQL vs. MySQL Community Edition

GreatSQL offers geographic tags, a new flow‑control algorithm, InnoDB parallel query and transaction‑lock optimizations, superior handling of network partitions, large‑transaction processing, node‑failure recovery, consistent read performance, increased MGR throughput, no data loss in multi‑write or single‑master failover, faster cluster startup, robust disk‑full handling, and resolves TCP self‑connect issues—advantages not present in MySQL Community Edition.

Building an MGR Cluster

docker‑compose Setup

Use docker‑compose to create a three‑node MGR cluster for testing.

Additional network configuration and password settings are included for immediate use.

<code>version: '3'
services:
  mgr1:
    image: greatsql/greatsql
    container_name: mgr1
    hostname: mgr1
    restart: unless-stopped
    environment:
      TZ: Asia/Shanghai
      MYSQL_ROOT_PASSWORD: root
      MYSQL_INIT_MGR: 1
      MYSQL_MGR_LOCAL: '172.27.0.2:33061'
      MYSQL_MGR_SEEDS: '172.27.0.2:33061,172.27.0.3:33061,172.27.0.4:33061'
    extra_hosts:
      - "mgr1:172.27.0.2"
      - "mgr2:172.27.0.3"
      - "mgr3:172.27.0.4"
    ports:
      - 3306:3306
    networks:
      mgr-net:
        ipv4_address: 172.27.0.2
  mgr2:
    image: greatsql/greatsql
    container_name: mgr2
    hostname: mgr2
    restart: unless-stopped
    depends_on:
      - "mgr1"
    environment:
      TZ: Asia/Shanghai
      MYSQL_ROOT_PASSWORD: root
      MYSQL_INIT_MGR: 1
      MYSQL_MGR_LOCAL: '172.27.0.3:33061'
      MYSQL_MGR_SEEDS: '172.27.0.2:33061,172.27.0.3:33061,172.27.0.4:33061'
    extra_hosts:
      - "mgr1:172.27.0.2"
      - "mgr2:172.27.0.3"
      - "mgr3:172.27.0.4"
    networks:
      mgr-net:
        ipv4_address: 172.27.0.3
  mgr3:
    image: greatsql/greatsql
    container_name: mgr3
    hostname: mgr3
    restart: unless-stopped
    depends_on:
      - "mgr2"
    environment:
      TZ: Asia/Shanghai
      MYSQL_ROOT_PASSWORD: root
      MYSQL_INIT_MGR: 1
      MYSQL_MGR_LOCAL: '172.27.0.4:33061'
      MYSQL_MGR_SEEDS: '172.27.0.2:33061,172.27.0.3:33061,172.27.0.4:33061'
    extra_hosts:
      - "mgr1:172.27.0.2"
      - "mgr2:172.27.0.3"
      - "mgr3:172.27.0.4"
    networks:
      mgr-net:
        ipv4_address: 172.27.0.4
networks:
  mgr-net:
    ipam:
      config:
        - subnet: 172.27.0.0/16</code>

Start MGR Service

Enter the PRIMARY node container (mgr1) and enable group replication.

<code>[root@greatsql]# docker exec -it mgr1 bash
[root@mgr1 /]# mysql
... 
[root@GreatSQL][(none)]> set global group_replication_bootstrap_group=ON;
[root@GreatSQL][(none)]> start group_replication;</code>

Start replication on the other two nodes.

<code>[root@greatsql]# docker exec -it mgr2 bash
[root@mgr2 /]# mysql
... 
[root@GreatSQL][(none)]> start group_replication;
Query OK, 0 rows affected (2.76 sec)

[root@greatsql]# docker exec -it mgr3 bash
[root@mgr3 /]# mysql
... 
[root@GreatSQL][(none)]> start group_replication;
Query OK, 0 rows affected (2.10 sec)</code>

Check MGR Status

<code>select * from performance_schema.replication_group_members;
+---------------------------+--------------------------------------+-------------+-------------+--------------+-------------+----------------+
| CHANNEL_NAME              | MEMBER_ID                           | MEMBER_HOST | MEMBER_PORT | MEMBER_STATE | MEMBER_ROLE | MEMBER_VERSION |
+---------------------------+--------------------------------------+-------------+-------------+--------------+-------------+----------------+
| group_replication_applier | 07455a01-0898-11ec-ac58-0242ac1b0002 | mgr1        | 3306        | ONLINE       | PRIMARY     | 8.0.25          |
| group_replication_applier | 124a583d-0898-11ec-bbb7-0242ac1b0003 | mgr2        | 3306        | ONLINE       | SECONDARY   | 8.0.25          |
| group_replication_applier | 1bf7df25-0898-11ec-9dd5-0242ac1b0004 | mgr3        | 3306        | ONLINE       | SECONDARY   | 8.0.25          |
+---------------------------+--------------------------------------+-------------+-------------+--------------+-------------+----------------+
3 rows in set (0.01 sec)</code>

PIG Microservice Integration

Modify the hosts mapping of the existing pig‑mysql service to point to GreatSQL.

The latest PIG master (3.3) is fully compatible with GreatSQL 8.0.25‑15.

Conclusion

The PIG microservice platform currently supports the following enterprise‑grade distributed databases:

GreatSQL 8.0.25‑15

TiDB 4.x

OceanBase 3.1

References

SpringCloud: https://spring.io/projects/spring-cloud

Service Mesh: https://istio.io/latest/about/service-mesh

Codis: https://github.com/CodisLabs/codis

GreatSQL background: https://www.greatdb.com

PIG: https://pig4cloud.com