MySQL集群搭建(6)-双主+keepalived高可用

双主 + keepalived 是一个比较简单的 MySQL 高可用架构，适用于中小 MySQL 集群，今天就说说怎么用 keepalived 做 MySQL 的高可用。

1 概述

1.1 keepalived 简介

简单地说，keepalived 就是通过管理 VIP 来实现机器的高可用的，在使用 keepalived 的情况下，只有一台服务器能够提供服务(通过 VIP 来实现)，当 Master 主机宕机后，VIP 会自动飘移到另一台服务器

keepalived 采用 Master/Slave 模式, 在 Master 上设置配置文件的 VIP，当 Master 宕机后，VIP 自动漂移到另一台 keepalived 服务器上

keepalived 可以用来做各种软件的高可用集群，它会一直检测服务器的状态，如果有一台服务器宕机，或工作出现故障，keepalived 将检测到，并将有故障的服务器从系统中剔除，同时使用其他服务器代替该服务器的工作，当服务器工作正常后 keepalived 自动将服务器加入到服务器群中。

1.2 keepalived 配合双主

keepalived 使用默认配置只能做到主机级别的高可用，但是我们的 MySQL 要做高可用至少要增加以下功能

• 能够检测 MySQL 服务状态
• 主节点 read_only=0 ，备节点 read_only=1
• 切换时，备节点要等待主节点同步完成

所以，keepalived 实现 MySQL 高可用需要使用自定义脚本来进行扩展

2 环境准备

2.1 数据库环境

操作前已经准备好了一套主主架构数据库，搭建方法参考 MySQL集群搭建(2)-主主从模式

节点信息

VIP 信息

Master 参考配置

[client]
port = 3306
default-character-set=utf8mb4
socket = /data/mysql_db/test_db/mysql.sock

[mysqld]
datadir = /data/mysql_db/test_db
basedir = /usr/local/mysql57
tmpdir = /tmp
socket = /data/mysql_db/test_db/mysql.sock
pid-file = /data/mysql_db/test_db/mysql.pid
skip-external-locking = 1
skip-name-resolve = 1
port = 3306
server_id = 2473306

default-storage-engine = InnoDB
character-set-server = utf8mb4
default_password_lifetime=0

auto_increment_offset = 1
auto_increment_increment = 2

#### log ####
log_timestamps=system
log_bin = /data/mysql_log/test_db/mysql-bin
log_bin_index = /data/mysql_log/test_db/mysql-bin.index
binlog_format = row
relay_log_recovery=ON
relay_log=/data/mysql_log/test_db/mysql-relay-bin
relay_log_index=/data/mysql_log/test_db/mysql-relay-bin.index
log_error = /data/mysql_log/test_db/mysql-error.log

#### replication ####
log_slave_updates = 1
replicate_wild_ignore_table = information_schema.%,performance_schema.%,sys.%

#### semi sync replication settings #####
plugin_dir=/usr/local/mysql57/lib/plugin
plugin_load = "rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so"
loose_rpl_semi_sync_master_enabled = 1
loose_rpl_semi_sync_slave_enabled = 1

Slave 参考配置

[client]
port = 3306
default-character-set=utf8mb4
socket = /data/mysql_db/test_db/mysql.sock

[mysqld]
datadir = /data/mysql_db/test_db
basedir = /usr/local/mysql57
tmpdir = /tmp
socket = /data/mysql_db/test_db/mysql.sock
pid-file = /data/mysql_db/test_db/mysql.pid
skip-external-locking = 1
skip-name-resolve = 1
port = 3306
server_id = 2483306

default-storage-engine = InnoDB
character-set-server = utf8mb4
default_password_lifetime=0

auto_increment_offset = 2
auto_increment_increment = 2

#### log ####
log_timestamps=system
log_bin = /data/mysql_log/test_db/mysql-bin
log_bin_index = /data/mysql_log/test_db/mysql-bin.index
binlog_format = row
relay_log_recovery=ON
relay_log=/data/mysql_log/test_db/mysql-relay-bin
relay_log_index=/data/mysql_log/test_db/mysql-relay-bin.index
log_error = /data/mysql_log/test_db/mysql-error.log

#### replication ####
log_slave_updates = 1
replicate_wild_ignore_table = information_schema.%,performance_schema.%,sys.%

#### semi sync replication settings #####
plugin_dir=/usr/local/mysql57/lib/plugin
plugin_load = "rpl_semi_sync_master=semisync_master.so;rpl_semi_sync_slave=semisync_slave.so"
loose_rpl_semi_sync_master_enabled = 1
loose_rpl_semi_sync_slave_enabled = 1

2.2 创建监控用的账号

- 由于是测试环境，账号密码设置比较随便
create user monitor@'localhost' identified by 'monitor';
grant all on *.* to monitor@'localhost';
flush privileges;

2.3 安装 keepalived

我们在 Master 和 Slave 上部署 keepalived

1). yum 安装

如果有对应的 yum 源，直接安装就可以了

yum install -y keepalived

2). 源码安装

下载安装包, 下载地址 keepalived , 使用 1.2.24 版本举例

# 安装依赖
yum install -y gcc popt-devel openssl openssl-devel libssl-dev libnl-devel popt-devel libnfnetlink-devel

# 下载包
wget http://www.keepalived.org/software/keepalived-1.2.24.tar.gz

# 解压安装
tar -xvz -f  keepalived-1.2.24.tar.gz
cd keepalived-1.2.24
./configure --prefix=/usr/local/keepalived
make && make install

cp /usr/local/keepalived/sbin/keepalived /usr/sbin/
cp /usr/local/keepalived/etc/rc.d/init.d/keepalived /etc/init.d/
cp /usr/local/keepalived/etc/sysconfig/keepalived /etc/sysconfig/
mkdir /etc/keepalived/
cp /usr/local/keepalived/etc/keepalived/keepalived.conf /etc/keepalived/

3 配置高可用

3.1 keepalived 配置

打开 /etc/keepalived/keepalived.conf 文件, 按照实际情况加上下面的配置

global_defs {
   router_id MYSQL_MM  # 标识
   vrrp_skip_check_adv_addr
   vrrp_strict        # 严格执行 VRRP 协议规范
   vrrp_garp_interval 0
   vrrp_gna_interval 0
}

vrrp_script check_mysql {
    script "/bin/sh /etc/keepalived/keepalived_mysql_check.sh"  # 检查脚本
    interval 10  # 检查周期
}

vrrp_instance MYSQL_MM {
    state BACKUP            # 都设为 BACKUP，避免起来后抢占
    interface eth0          # 网卡名称，根据实际情况填写
    virtual_router_id 243   # 用来区分 VRRP 组播的标记，取值 0-255
    priority 100
    advert_int 1
    nopreempt               # 设为非抢占
    authentication {
        auth_type PASS
        auth_pass 1111
    }

    # Master 节点可以注释掉下面语句，防止启动 keepalived 的时候执行脚本
    notify_master "/bin/sh /etc/keepalived/keepalived_mysql_start.sh"  # 变为 MASTER 时执行

    virtual_ipaddress {
        10.0.0.237
    }

    # Slave 节点可以注释下面检查脚本，Slave 没有必要一直检查
    track_script {
        check_mysql
    }
}

3.2 配置检查脚本

打开 /etc/keepalived/keepalived_mysql_check.sh ， 写入检测脚本

#!/bin/sh
# @Author: chengqm
# MySQL 检测脚本
MyPath=$(cd $(dirname $0); pwd)
cd $MyPath

ThisTime=`date '+%F %T'`
log_file='/var/log/keepalived_mysql.log'

# MySQL 连接方式，根据实际情况调整
export MYSQL_PWD='monitor'
MYSQL_USER='monitor'
MYSQL_SOCKET="/data/mysql_db/test_db/mysql.sock"
mysql_connect="mysql -u${MYSQL_USER} -S${MYSQL_SOCKET} "

# 美化输出
function techo() {
    message=$1
    message_level=$2
    if [ -e $message_level ];then
        message_level='info'
    fi
    echo "`date '+%F %T'` - [${message_level}] $message" >> $log_file
}

# 检查函数, 正常返回 0
function check {
    ret=`$mysql_connect -N -e 'select 1 as value'`
    if [ $? -ne 0 ] || [ $ret -ne '1' ];then
        return 1
    else
        return 0
    fi
}

function read_only {
    param=$1
    $mysql_connect -e "set global read_only = ${param}"
    techo "设置是否只读 read_only ${param}"
}

# 失效转移
function failover {
    techo "开始执行失效转移"
    # 1. 停止 keepalived
    killall keepalived

    # 2. 如果还能执行的话，设为 read_only
    read_only 1

    if [ $? -eq 0 ];then
        # 3. 如果还能执行，kill 所有的连接
        $mysql_connect -e "select concat('KILL ',id,';') from information_schema.processlist where user!='root' AND db is not null into outfile '/tmp/kill.txt.${ThisTime}';"
        if [ $? -eq 0 ];then
            $mysql_connect -e "source /tmp/kill.txt.${ThisTime};"
        fi
    fi

    # 4. 其他操作，比如说自动关机

    techo "失效转移执行成功，当前数据库关闭访问"
}

# 有问题检查 4 次
for ((i=1; i<=4; i ++))  
do  
    check
    if [ $? -eq 0 ];then
        techo "MySQL is ok"
        # 正常退出脚本
        exit 0
    else
        techo "Connection failed $i time(s)"
        sleep 1
    fi
done

techo '无法连接当前数据库'

# 失效转移
failover

注意：脚本没有经过严格测试，需要根据实际情况调整

3.3 配置提升为 Master 时执行的脚本

打开 /bin/sh /etc/keepalived/keepalived_mysql_start.sh" , 写入脚本内容

#!/bin/sh
# @Author: chengqm
# keepalived 变为 Master 时执行
MyPath=$(cd $(dirname $0); pwd)
cd $MyPath

ThisTime=`date '+%F %T'`
log_file='/var/log/keepalived_mysql.log'

# MySQL 连接方式，根据实际情况调整
export MYSQL_PWD='monitor'
MYSQL_USER='monitor'
MYSQL_SOCKET="/data/mysql_db/test_db/mysql.sock"
mysql_connect="mysql -u${MYSQL_USER} -S${MYSQL_SOCKET} "

# 美化输出
function techo() {
    message=$1
    message_level=$2
    if [ -e $message_level ];then
        message_level='info'
    fi
    echo "`date '+%F %T'` - [${message_level}] $message" >> $log_file
}

# 检查函数, 正常返回 0
function check {
    ret=`$mysql_connect -N -e 'select 1 as value'`
    if [ $? -ne 0 ] || [ $ret -ne '1' ];then
        return 1
    else
        return 0
    fi
}

# 获取 slave status 的信息
function slave_info() {
    tmp_file=/tmp/slave_info.tmp
    $mysql_connect -e 'show slave status\G' > /tmp/slave_info.tmp
    slave_sql=`grep 'Slave_SQL_Running:' $tmp_file | sed 's/\s*//g' | tr "A-Z" "a-z"  | awk -F":" '{print $2}'`
    seconds_behind_master=`grep 'Seconds_Behind_Master:' $tmp_file | sed 's/\s*//g' | tr "A-Z" "a-z"  | awk -F":" '{print $2}'`

    master_log_file=`grep 'Master_Log_File:' $tmp_file | head -1 | sed 's/\s*//g' | tr "A-Z" "a-z"  | awk -F":" '{print $2}'`
    master_log_pos=`grep 'Read_Master_Log_Pos:' $tmp_file | sed 's/\s*//g' | tr "A-Z" "a-z"  | awk -F":" '{print $2}'`

    relay_master_log_file=`grep 'Relay_Master_Log_File:' $tmp_file | sed 's/\s*//g' | tr "A-Z" "a-z"  | awk -F":" '{print $2}'`
    exec_master_log_pos=`grep 'Exec_Master_Log_Pos:' $tmp_file | sed 's/\s*//g' | tr "A-Z" "a-z"  | awk -F":" '{print $2}'`

}

# 设置是否可读
function read_only {
    param=$1
    $mysql_connect -e "set global read_only = ${param}"
    techo "设置是否只读 read_only ${param}"
}

# 处理数据同步
function sync_master_log() {
    # 如果是数据一致性优先，等待同步完毕。如果是服务可用性优先，可以注销下面的代码
    slave_info
    if [ $slave_sql == "yes" ];then
        techo "当前同步位置 Master ${master_log_file} ${master_log_pos}"
        techo "等待同步到 Master ${master_log_file} ${master_log_pos}"
        $mysql_connect -e "select master_pos_wait('$master_log_file', $master_log_pos);" > /dev/null
        techo "同步完毕"
    fi
}

techo "当前数据库提升为主库"

check
if [ $? -ne 0 ];then
    techo "无法连接当前数据库"
    exit 1
fi

# 等待同步
sync_master_log

# 设为可写
read_only 0

注意：脚本没有经过严格测试，需要根据实际情况调整

3.4 启动 keepalived

由于配置了 BACKUP 模式，所以两个 keepalived 先起来的是主，先后在主备节点执行

/etc/init.d/keepalived start

检查 /var/log/message 日志，确认 keepalived 没有报错

检查 Master IP 状态, 确认设置了 VIP

[root@cluster01 shell]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN 
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether fa:16:3e:de:80:33 brd ff:ff:ff:ff:ff:ff
    inet 10.0.0.247/16 brd 10.0.255.255 scope global eth0
    inet 10.0.0.237/32 scope global eth0
    inet6 fe80::f816:3eff:fede:8033/64 scope link 
       valid_lft forever preferred_lft forever

检查 MySQL 检测脚本执行情况，确认正常运行

[root@cluster01 ~]# tail -f /var/log/keepalived_mysql.log 
...
2019-01-28 15:04:18 - [info] MySQL is ok
2019-01-28 15:04:28 - [info] MySQL is ok

4 失效转移测试

在 mytest 库里新建 nowdate 测试表，只有 id 和 ctime 字段，然后每秒插入一条数据

[root@cluster03 ~]# while true; do date;mysql -h10.0.0.237 -P3306 -umytest -e 'use mytest;insert into nowdate values (null, now());'; sleep 1;done
Mon Jan 28 15:04:26 CST 2019
Mon Jan 28 15:04:27 CST 2019
...

kill 掉 Master 进程

killall mysqld

查看旧 Master 日志

2019-01-28 15:04:48 - [info] MySQL is ok
2019-01-28 15:04:58 - [info] Connection failed 1 time(s)
2019-01-28 15:04:59 - [info] Connection failed 2 time(s)
2019-01-28 15:05:00 - [info] Connection failed 3 time(s)
2019-01-28 15:05:01 - [info] Connection failed 4 time(s)
2019-01-28 15:05:02 - [info] 无法连接当前数据库
2019-01-28 15:05:02 - [info] 开始执行失效转移
2019-01-28 15:05:02 - [info] 设置是否只读 read_only 1
2019-01-28 15:05:02 - [info] 失效转移执行成功，当前数据库关闭访问

查看新 Master 日志

2019-01-28 15:05:04 - [info] 当前数据库提升为主库
2019-01-28 15:05:04 - [info] 当前同步位置 Master mysql-bin.000015 32338
2019-01-28 15:05:04 - [info] 等待同步到 Master mysql-bin.000015 32338
2019-01-28 15:05:04 - [info] 同步完毕
2019-01-28 15:05:04 - [info] 设置是否只读 read_only 0
2019-01-28 15:05:05 - [info] MySQL is ok

查看新 Master IP，确认 VIP 已经飘过来了

[root@cluster02 ~]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN 
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether fa:16:3e:66:7e:e8 brd ff:ff:ff:ff:ff:ff
    inet 10.0.0.248/16 brd 10.0.255.255 scope global eth0
    inet 10.0.0.237/32 scope global eth0
    inet6 fe80::f816:3eff:fe66:7ee8/64 scope link 
       valid_lft forever preferred_lft forever

查看插入数据执行情况，大概有 12 秒是不可用的

Mon Jan 28 15:04:51 CST 2019
ERROR 2003 (HY000): Can't connect to MySQL server on '10.0.0.237' (111)
Mon Jan 28 15:04:52 CST 2019
ERROR 2003 (HY000): Can't connect to MySQL server on '10.0.0.237' (111)
Mon Jan 28 15:04:53 CST 2019
ERROR 2003 (HY000): Can't connect to MySQL server on '10.0.0.237' (111)
Mon Jan 28 15:04:54 CST 2019
ERROR 2003 (HY000): Can't connect to MySQL server on '10.0.0.237' (111)
Mon Jan 28 15:04:55 CST 2019
ERROR 2003 (HY000): Can't connect to MySQL server on '10.0.0.237' (111)
Mon Jan 28 15:04:56 CST 2019
ERROR 2003 (HY000): Can't connect to MySQL server on '10.0.0.237' (111)
Mon Jan 28 15:04:57 CST 2019
ERROR 2003 (HY000): Can't connect to MySQL server on '10.0.0.237' (111)
Mon Jan 28 15:04:58 CST 2019
ERROR 2003 (HY000): Can't connect to MySQL server on '10.0.0.237' (111)
Mon Jan 28 15:05:00 CST 2019
ERROR 2003 (HY000): Can't connect to MySQL server on '10.0.0.237' (111)
Mon Jan 28 15:05:01 CST 2019
ERROR 2003 (HY000): Can't connect to MySQL server on '10.0.0.237' (111)
Mon Jan 28 15:05:02 CST 2019
ERROR 2003 (HY000): Can't connect to MySQL server on '10.0.0.237' (111)
Mon Jan 28 15:05:03 CST 2019

失效切换成功

5 总结

使用双主 + keepalived 的优点是部署简单，双主加半同步情况下，理论上不会丢数据，适用于中小型 MySQL 集群。缺点也比较明显，就是增加从节点的情况下，从节点不会主动切换同步对象，而且脚本需要自己实现，有一定风险。