keepalived mysql 主从复制 容器实现(失败)

由于keepalived本身机制问题 无法实现容器层的单机部署两个keepalived及两个mysql主从复制，故障后自动漂移至另一个可用的节点的效果，原因在末尾
数据库主从复制配置部分依然可用

准备

创建网络

docker network create \--subnet=192.168.10.0/24 \--gateway=192.168.10.1 \--driver=bridge docker-mysql-ha

启动脚本

services:mysql-master:container_name: mysql-masterhostname: mysql-masterimage: mysql:8.0networks:docker-mysql-ha:ipv4_address: 192.168.10.11environment:TZ: Asia/ShanghaiMYSQL_ROOT_PASSWORD: 20240510MYSQL_USER: kd1MYSQL_PASSWORD: 20240511volumes:- /etc/localtime:/etc/localtime:ro- /data/dockerfiles/mysql-master/mysql_data:/var/lib/mysql- /data/dockerfiles/mysql-master/conf/my.cnf:/etc/my.cnf- /data/dockerfiles/mysql-master/mysqld:/var/run/mysqld- /data/dockerfiles/mysql-master/log:/var/log/mysqlrestart: unless-stoppedmysql-slave:container_name: mysql-slavehostname: mysql-slaveimage: mysql:8.0networks:docker-mysql-ha:ipv4_address: 192.168.10.12environment:TZ: Asia/ShanghaiMYSQL_ROOT_PASSWORD: 20240510MYSQL_USER: kd1MYSQL_PASSWORD: 20240511volumes:- /etc/localtime:/etc/localtime:ro- /data/dockerfiles/mysql-slave/mysql_data:/var/lib/mysql- /data/dockerfiles/mysql-slave/conf/my.cnf:/etc/my.cnf- /data/dockerfiles/mysql-slave/mysqld:/var/run/mysqld- /data/dockerfiles/mysql-slave/log:/var/log/mysqlrestart: unless-stoppednetworks:docker-mysql-ha:external: true

创建持久化存储目录

mkdir -p /data/dockerfiles/mysql-master/conf/
mkdir -p /data/dockerfiles/mysql-slave/conf/

放入配置文件

主节点配置
my.cnf

# For advice on how to change settings please see
# http://dev.mysql.com/doc/refman/8.0/en/server-configuration-defaults.html[mysqld]
#
# Remove leading # and set to the amount of RAM for the most important data
# cache in MySQL. Start at 70% of total RAM for dedicated server, else 10%.
# innodb_buffer_pool_size = 128M
#
# Remove leading # to turn on a very important data integrity option: logging
# changes to the binary log between backups.
# log_bin
#
# Remove leading # to set options mainly useful for reporting servers.
# The server defaults are faster for transactions and fast SELECTs.
# Adjust sizes as needed, experiment to find the optimal values.
# join_buffer_size = 128M
# sort_buffer_size = 2M
# read_rnd_buffer_size = 2M# Remove leading # to revert to previous value for default_authentication_plugin,
# this will increase compatibility with older clients. For background, see:
# https://dev.mysql.com/doc/refman/8.0/en/server-system-variables.html#sysvar_default_authentication_plugin
# default-authentication-plugin=mysql_native_password# 基本路径设置
datadir=/var/lib/mysql
socket=/var/run/mysqld/mysqld.sock
user=mysql
pid-file=/var/run/mysqld/mysqld.pid
secure-file-priv=/var/lib/mysql-files# 网络与连接
skip-host-cache
skip-name-resolve
max_connections=200
wait_timeout=28800
interactive_timeout=28800# 字符集与排序规则
character-set-server=utf8mb4
collation-server=utf8mb4_unicode_ci
init_connect='SET NAMES utf8mb4'# 设置服务器时区（建议与宿主机一致）
default-time-zone='+08:00'# 二进制日志（推荐开启，用于数据恢复/主从复制）
server-id=1
binlog_format=ROW
log_bin=/var/log/mysql/mysql-bin.log
binlog_expire_logs_seconds=604800  # 保留7天
sync_binlog=1# 慢查询日志（调优关键）
slow_query_log=1
slow_query_log_file=/var/log/mysql/mysql-slow.log
long_query_time=1
log_queries_not_using_indexes=1# 错误日志
# log_error=/var/log/mysql/mysql-error.log# InnoDB 引擎优化
innodb_buffer_pool_size=512M
innodb_redo_log_capacity=256M
innodb_file_per_table=1
innodb_flush_log_at_trx_commit=1# 缓存优化
table_open_cache=400
tmp_table_size=64M
max_heap_table_size=64M
sort_buffer_size=4M
read_buffer_size=2M
join_buffer_size=4M[client]
socket=/var/run/mysqld/mysqld.sock
default-character-set=utf8mb4[mysqldump]
default-character-set=utf8mb4
quick
max_allowed_packet=64M!includedir /etc/mysql/conf.d/

从节点配置
my.cnf

# For advice on how to change settings please see
# http://dev.mysql.com/doc/refman/8.0/en/server-configuration-defaults.html[mysqld]
#
# Remove leading # and set to the amount of RAM for the most important data
# cache in MySQL. Start at 70% of total RAM for dedicated server, else 10%.
# innodb_buffer_pool_size = 128M
#
# Remove leading # to turn on a very important data integrity option: logging
# changes to the binary log between backups.
# log_bin
#
# Remove leading # to set options mainly useful for reporting servers.
# The server defaults are faster for transactions and fast SELECTs.
# Adjust sizes as needed, experiment to find the optimal values.
# join_buffer_size = 128M
# sort_buffer_size = 2M
# read_rnd_buffer_size = 2M# Remove leading # to revert to previous value for default_authentication_plugin,
# this will increase compatibility with older clients. For background, see:
# https://dev.mysql.com/doc/refman/8.0/en/server-system-variables.html#sysvar_default_authentication_plugin
# default-authentication-plugin=mysql_native_password# 基本路径设置
datadir=/var/lib/mysql
socket=/var/run/mysqld/mysqld.sock
user=mysql
pid-file=/var/run/mysqld/mysqld.pid
secure-file-priv=/var/lib/mysql-files# 网络与连接
skip-host-cache
skip-name-resolve
max_connections=200
wait_timeout=28800
interactive_timeout=28800# 字符集与排序规则
character-set-server=utf8mb4
collation-server=utf8mb4_unicode_ci
init_connect='SET NAMES utf8mb4'# 设置服务器时区（建议与宿主机一致）
default-time-zone='+08:00'# 二进制日志（推荐开启，用于数据恢复/主从复制）
server-id=2
relay-log=/var/lib/mysql/relay-log
binlog_format=ROW
log_bin=/var/log/mysql/mysql-bin.log
binlog_expire_logs_seconds=604800  # 保留7天
sync_binlog=1# 慢查询日志（调优关键）
slow_query_log=1
slow_query_log_file=/var/log/mysql/mysql-slow.log
long_query_time=1
log_queries_not_using_indexes=1# 错误日志
# log_error=/var/log/mysql/mysql-error.log# InnoDB 引擎优化
innodb_buffer_pool_size=512M
innodb_redo_log_capacity=256M
innodb_file_per_table=1
innodb_flush_log_at_trx_commit=1# 缓存优化
table_open_cache=400
tmp_table_size=64M
max_heap_table_size=64M
sort_buffer_size=4M
read_buffer_size=2M
join_buffer_size=4M[client]
socket=/var/run/mysqld/mysqld.sock
default-character-set=utf8mb4[mysqldump]
default-character-set=utf8mb4
quick
max_allowed_packet=64M!includedir /etc/mysql/conf.d/

启动容器

docker compose -f /data/dockercompose/docker-compose-mysql-ha.yml up -d

Mysql主从复制配置

主库配置

docker exec -it mysql-master mysql -uroot -p20240510# 创建一个用户名叫 replicator 的新用户。 '%' 表示允许从任意 IP 地址连接。 账号的密码是 replica123。
mysql> CREATE USER 'replicator'@'%' IDENTIFIED BY 'replica123';# 给 replicator 用户授予 REPLICATION SLAVE 权限。 这个权限允许该用户作为从库连接主库，进行二进制日志的复制。 *.* 表示数据库和表全部适用。
mysql> GRANT REPLICATION SLAVE ON *.* TO 'replicator'@'%';# 刷新权限表：
mysql> FLUSH PRIVILEGES;# 查看当前 binlog 状态 记住 File 和 Position，后面从库要用。
mysql> SHOW MASTER STATUS;+------------------+----------+--------------+------------------+-------------------+
| File             | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin.000005 |      868 |              |                  |                   |
+------------------+----------+--------------+------------------+-------------------+
1 row in set (0.00 sec)

从库配置

docker exec -it mysql-slave mysql -uroot -p20240510# 从库中设置主库连接信息
mysql> CHANGE MASTER TOMASTER_HOST='192.168.10.11',MASTER_PORT=3306,MASTER_USER='replicator',MASTER_PASSWORD='replica123',MASTER_LOG_FILE='mysql-bin.000005',  -- 此处替换为主库 SHOW MASTER STATUS 的结果MASTER_LOG_POS=868,                 -- 同上GET_MASTER_PUBLIC_KEY = 1;# 启动复制
mysql> START SLAVE;# 查看复制状态
mysql> SHOW SLAVE STATUS\G确认以下两项均为 Yes：
Slave_IO_Running: Yes
Slave_SQL_Running: Yes

示例输出：
mysql> SHOW SLAVE STATUS\G
*************************** 1. row ***************************Slave_IO_State: Waiting for source to send eventMaster_Host: 192.168.10.11Master_User: replicatorMaster_Port: 3306Connect_Retry: 60Master_Log_File: mysql-bin.000005Read_Master_Log_Pos: 868Relay_Log_File: relay-log.000002Relay_Log_Pos: 326Relay_Master_Log_File: mysql-bin.000005Slave_IO_Running: Yes              -- 此处应为 YesSlave_SQL_Running: Yes              -- 此处应为 YesReplicate_Do_DB: Replicate_Ignore_DB: Replicate_Do_Table: Replicate_Ignore_Table: Replicate_Wild_Do_Table: Replicate_Wild_Ignore_Table: Last_Errno: 0Last_Error: Skip_Counter: 0Exec_Master_Log_Pos: 868Relay_Log_Space: 530Until_Condition: NoneUntil_Log_File: Until_Log_Pos: 0Master_SSL_Allowed: NoMaster_SSL_CA_File: Master_SSL_CA_Path: Master_SSL_Cert: Master_SSL_Cipher: Master_SSL_Key: Seconds_Behind_Master: 0
Master_SSL_Verify_Server_Cert: NoLast_IO_Errno: 0Last_IO_Error: Last_SQL_Errno: 0Last_SQL_Error: Replicate_Ignore_Server_Ids: Master_Server_Id: 1Master_UUID: b8b661ab-6d15-11f0-be67-0242c0a80a0bMaster_Info_File: mysql.slave_master_infoSQL_Delay: 0SQL_Remaining_Delay: NULLSlave_SQL_Running_State: Replica has read all relay log; waiting for more updatesMaster_Retry_Count: 86400Master_Bind: Last_IO_Error_Timestamp: Last_SQL_Error_Timestamp: Master_SSL_Crl: Master_SSL_Crlpath: Retrieved_Gtid_Set: Executed_Gtid_Set: Auto_Position: 0Replicate_Rewrite_DB: Channel_Name: Master_TLS_Version: Master_public_key_path: Get_master_public_key: 1Network_Namespace: 
1 row in set, 1 warning (0.00 sec)

验证同步

• 主库操作

CREATE DATABASE test_sync;
USE test_sync;
CREATE TABLE users (id INT PRIMARY KEY, name VARCHAR(100));
INSERT INTO users VALUES (1, 'Alice');

• 从库验证

SELECT * FROM test_sync.users;# 示例输出
+----+-------+
| id | name  |
+----+-------+
|  1 | Alice |
+----+-------+
1 row in set (0.00 sec)

更改主从配置

• 主库

docker exec -it mysql-master mysql -uroot -p20240510-- 确保之前创建用于主从配置的用户存在
mysql> SELECT user, host FROM mysql.user WHERE user = 'replicator';
+------------+------+
| user       | host |
+------------+------+
| replicator | %    |
+------------+------+
1 row in set (0.00 sec)-- 再次查看当前 binlog 状态 记住 File 和 Position，后面从库要用。
mysql> SHOW MASTER STATUS;
+------------------+----------+--------------+------------------+-------------------+
| File             | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin.000011 |      756 |              |                  |                   |
+------------------+----------+--------------+------------------+-------------------+
1 row in set (0.00 sec)

• 从库

docker exec -it mysql-slave mysql -uroot -p20240510-- 停止之前配置的复制的 IO 线程
STOP REPLICA IO_THREAD;-- 从库中设置主库连接信息
CHANGE MASTER TOMASTER_HOST='192.168.0.8',MASTER_PORT=3306,MASTER_USER='replicator',MASTER_PASSWORD='replica123',MASTER_LOG_FILE='mysql-bin.000011', MASTER_LOG_POS=456,GET_MASTER_PUBLIC_KEY = 1;-- 启动复制
mysql> START SLAVE;-- 查看复制状态
mysql> SHOW SLAVE STATUS\G确认以下两项均为 Yes：
Slave_IO_Running: Yes
Slave_SQL_Running: Yes

验证同步

• 主库操作

USE test_sync;
INSERT INTO users VALUES (5, 'Slice');

• 从库验证

SELECT * FROM test_sync.users;# 示例输出
+----+-------+
| id | name  |
+----+-------+
|  1 | Alice |
|  5 | Slice |
+----+-------+
2 rows in set (0.00 sec)

取消从库主从复制

-- 停止复制线程
STOP SLAVE;-- 重置复制配置（清除复制相关信息）
RESET SLAVE ALL;-- 检查当前是否启用了 read_only; ON 表示开启; OFF 表示关闭
SHOW VARIABLES LIKE 'read_only';-- 可选）如果你想让从库完全变成普通库，且保证可写，执行：
SET GLOBAL read_only = OFF;-- （可选）检查复制状态，确认已经取消; 如果显示为空或报错，说明复制已取消。
SHOW SHOW SLAVE STATUS\G

安装 Keepalived 容器（两个）

• 需要部署两个 Keepalived 容器：分别绑定到 mysql-master 和 mysql-slave 上。
  Keepalived 容器共享主机的网络命名空间（host 模式），以便绑定 VIP：

# 创建配置目录
mkdir -p /data/dockerfiles/keepalived/master
mkdir -p /data/dockerfiles/keepalived/slave

编写 Keepalived 配置

注意：interface 应该写 docker-mysql-ha 也就是192.168.10.x 网络对应的网卡名

• master 端：
  文件：/data/dockerfiles/keepalived/master/keepalived.conf

vrrp_script check_mysql {script "/usr/local/etc/keepalived/check_mysql.sh 192.168.10.11"interval 2weight -30
}vrrp_instance VI_1 {state MASTERinterface eth0virtual_router_id 51priority 150advert_int 1authentication {auth_type PASSauth_pass 1234}virtual_ipaddress {172.21.147.250}track_script {check_mysql}
}

• 脚本：/data/dockerfiles/keepalived/master/check_mysql.sh

#!/bin/bash
CHECK_HOST=$1
CHECK_PORT=3306timeout 1 bash -c "echo > /dev/tcp/${CHECK_HOST}/${CHECK_PORT}"
if [ $? -eq 0 ]; thenexit 0
elseexit 1
fi

• slave 端：
  文件：/data/dockerfiles/keepalived/slave/keepalived.conf

vrrp_script check_mysql {script "/usr/local/etc/keepalived/check_mysql.sh 192.168.10.12"interval 2weight -30
}vrrp_instance VI_1 {state BACKUPinterface eth0virtual_router_id 51priority 100advert_int 1authentication {auth_type PASSauth_pass 1234}virtual_ipaddress {172.21.147.250}track_script {check_mysql}
}

• 脚本：/data/dockerfiles/keepalived/slave/check_mysql.sh

#!/bin/bash
CHECK_HOST=$1
CHECK_PORT=3306timeout 1 bash -c "echo > /dev/tcp/${CHECK_HOST}/${CHECK_PORT}"
if [ $? -eq 0 ]; thenexit 0
elseexit 1
fi

• 别忘了给脚本加执行权限：

chmod +x /data/dockerfiles/keepalived/*/check_mysql.sh

启动脚本

确保容器使用 host 网络（否则无法绑定 VIP）：

version: "3.8"services:keepalived-master:image: osixia/keepalived:2.0.20container_name: keepalived-masternetwork_mode: hostcap_add:- NET_ADMIN- NET_RAW- NET_BROADCASTvolumes:- /data/dockerfiles/keepalived/master/keepalived.conf:/usr/local/etc/keepalived/keepalived.conf:ro- /data/dockerfiles/keepalived/master/check_mysql.sh:/usr/local/etc/keepalived/check_mysql.sh:rorestart: alwayskeepalived-slave:image: osixia/keepalived:2.0.20container_name: keepalived-slavenetwork_mode: hostcap_add:- NET_ADMIN- NET_RAW- NET_BROADCASTvolumes:- /data/dockerfiles/keepalived/slave/keepalived.conf:/usr/local/etc/keepalived/keepalived.conf:ro- /data/dockerfiles/keepalived/slave/check_mysql.sh:/usr/local/etc/keepalived/check_mysql.sh:rorestart: always

原因

一、致命架构缺陷

‌VIP与宿主机IP跨子网问题‌

• VIP(192.168.1.100/24)与宿主机IP(192.168.10.25/24)处于不同子网
• 二层网络无法直接通信，ARP广播请求无法到达
• 违反网络基础原理：同一物理接口不能承载跨子网IP

‌路由黑洞问题‌

• 192.168.1.0/24网段在路由器无路由条目
• 返回数据包因非对称路由被丢弃

‌ARP代理失效‌

• 宿主机无法响应192.168.1.0/24网段的ARP请求
• 客户端持续发送ARP请求无响应

二、技术方案矛盾点

方案要素	矛盾点	后果
单机VIP漂移	单节点无需VIP漂移	增加无效复杂度
端口重定向(DNAT)	与VIP功能重叠	流量路径混乱
Keepalived监控	单点监控无意义	无法实现真正HA
跨子网VIP	违反RFC网络标准	连通性不可达

三、网络通信原理冲突

‌OSI模型违反‌

• 三层IP(192.168.1.100)绑定在二层接口(eth0)
• 但eth0已绑定192.168.10.0/24子网
• 违反"一个物理接口不同时承载多个逻辑子网"原则

ARP协议限制‌

# 客户端ARP请求
ARP who-has 192.168.1.100 tell 192.168.10.24# 宿主机响应
# 因IP不属于接口子网，内核自动丢弃该请求！