keepalived 做mysql的主从切换
适用于mysql master-slave的主从架构
一、过程简介：
1、通过keepalived配置VIP高可用，keepalived均设置为BACKUP ,nopreempt非抢占模式。 2、master上监控mysql 3306端口的状态，当检测到3306 端口停止后，停止keepalived，vip自动转移到slave上。 3、slave获取到vip升级为master后，执行changemasterdb脚本，将本地mysql库升级为master库，同时将其他从库的主库修改为本库的从库。
二、主从架构图
2.1、切换前的
2.2、切换后的
三、主要配置文件
主要配置文件及作用说明
3.1 keepalived.conf.master 用于当前为master的数据库上的keepalived的配置文件
[root@t156 keepalived]# more keepalived.conf.master
! Configuration File for Keepalived
! ---------------------------------------------------------------------------
! GLOBAL
! ---------------------------------------------------------------------------
global_defs {
! this is who emails will go to on alerts
notification_email {
wanghengzhi@hw801.com
! add a few more email addresses here if you would like
}
notification_email_from wanghengzhi@hw801.com
! mail relay server
smtp_server 127.0.0.1
smtp_connect_timeout 30
! each load balancer should have a different ID
! this will be used in SMTP alerts, so you should make
! each router easily identifiable
router_id LVS_170
vrrp_mcast_group4 224.0.0.18
lvs_sync_daemon eth1 VI1_LVS_DB
script_user root
}
vrrp_script check_mysql {
script "/etc/keepalived/mysql_check.sh"
interval 10
}
vrrp_instance VI1_LVS_DB {
!state MASTER
state BACKUP
interface eth1
track_interface {
eth1
}
! interface to run LVS sync daemon on
! lvs_sync_daemon_interface eth1
!mcast_src_ip 192.168.1.156
! each virtual router id must be unique per instance name!
virtual_router_id 170
! MASTER and BACKUP state are determined by the priority
! even if you specify MASTER as the state, the state will
! be voted on by priority (so if your state is MASTER but your
! priority is lower than the router with BACKUP, you will lose
! the MASTER state)
! I make it a habit to set priorities at least 50 points apart
! note that a lower number is lesser priority - lower gets less vote
priority 100
! how often should we vote, in seconds?
advert_int 1
! send an alert when this instance changes state from MASTER to BACKUP
smtp_alert
! this authentication is for syncing between failover servers
! keepalived supports PASS, which is simple password
! authentication or AH, which is the IPSec authentication header.
! Don't use AH yet as many people have reported problems with it
authentication {
auth_type PASS
auth_pass 111111
}
! these are the IP addresses that keepalived will setup on this
! machine. Later in the config we will specify which real
! servers are behind these IPs without this block, keepalived
! will not setup and takedown any IP addresses
virtual_ipaddress {
192.168.1.170/24 dev eth1
}
nopreempt
!preempt_delay 2
track_script {
check_mysql
}
!notify_master "/etc/keepalived/changemasterdb.sh"
}
注：
state BACKUP #state 全部为BACKUP
priority 100 #当前是master的优先级高
vrrp_script check_mysql {
script "/etc/keepalived/mysql_check.sh" #vrrp_script 检测脚本
interval 10
}
nopreempt # vip 设置为非抢占模式
track_script {
check_mysql # 调用 vrrp_script检测脚本
}
3.2 keepalived.conf.slave 用于当前为slave的数据库上的keepalived的配置文件
[root@t168 keepalived]# more keepalived.conf.backup
! Configuration File for Keepalived
! ---------------------------------------------------------------------------
! GLOBAL
! ---------------------------------------------------------------------------
global_defs {
! this is who emails will go to on alerts
notification_email {
wanghengzhi@hw801.com
! add a few more email addresses here if you would like
}
notification_email_from wanghengzhi@hw801.com
! mail relay server
smtp_server 127.0.0.1
smtp_connect_timeout 30
! each load balancer should have a different ID
! this will be used in SMTP alerts, so you should make
! each router easily identifiable
router_id LVS_170
vrrp_mcast_group4 224.0.0.18
lvs_sync_daemon eth1 VI1_LVS_DB
script_user root
}
vrrp_instance VI1_LVS_DB {
state BACKUP
interface eth1
track_interface {
eth1
}
! interface to run LVS sync daemon on
! lvs_sync_daemon_interface eth1
!mcast_src_ip 192.168.1.168
! each virtual router id must be unique per instance name!
virtual_router_id 170
! MASTER and BACKUP state are determined by the priority
! even if you specify MASTER as the state, the state will
! be voted on by priority (so if your state is MASTER but your
! priority is lower than the router with BACKUP, you will lose
! the MASTER state)
! I make it a habit to set priorities at least 50 points apart
! note that a lower number is lesser priority - lower gets less vote
priority 90
! how often should we vote, in seconds?
advert_int 1
! send an alert when this instance changes state from MASTER to BACKUP
smtp_alert
! this authentication is for syncing between failover servers
! keepalived supports PASS, which is simple password
! authentication or AH, which is the IPSec authentication header.
! Don't use AH yet as many people have reported problems with it
authentication {
auth_type PASS
auth_pass 111111
}
! these are the IP addresses that keepalived will setup on this
! machine. Later in the config we will specify which real
! servers are behind these IPs without this block, keepalived
! will not setup and takedown any IP addresses
virtual_ipaddress {
192.168.1.170/24 dev eth1
}
nopreempt
notify_master "/etc/keepalived/changemasterdb.sh"
}
注：
state BACKUP #state 全部为BACKUP priority 90 #当前是master的优先级高 nopreempt # vip 设置为非抢占模式 notify_master "/etc/keepalived/changemasterdb.sh" # 切换为master后执行的脚本
3.3 mysql_check.sh 用于主库上mysql 3306端口的检测
[root@t168 keepalived]# more mysql_check.sh
#!/bin/bash
set -o nounset
#@Author : wanghz
#@Time : 2021/9/1 12:05
# define restricted path
PATH="/bin:/usr/bin:/sbin:/usr/sbin"
# adirname - return absolute dirname of given file
adirname() { odir=`pwd`; cd `dirname $1`; pwd; cd "${odir}"; }
PSW="123456"
PORT='3306'
MYSQL_BIN=`which mysql`
MYSQL_MASTER_BIN="${MYSQL_BIN} -uroot -p${PSW} -S /tmp/mysql${PORT}.sock"
count=1
while true
do
${MYSQL_MASTER_BIN} -e "show status\G;" >/dev/null 2>&1
Status=$?
ps aux|grep mysqld|grep -v grep >/dev/null 2>&1
Grep=$?
if [ ${Status} = 0 ] && [ ${Grep} = 0 ]
then
exit 0
else
if [ ${Status} = 1 ] && [ ${Grep} = 0 ]
then
exit 0
else
if [ ${count} -gt 5 ]
then
echo ${count}
break
fi
let count+=1
continue
fi
fi
done
`which service` keepalived stop
3.4 changemasterdb.sh 用于到vip 切换到从库上，进行的一系列操作
[root@t168 keepalived]# more changemasterdb.sh
#!/bin/bash
set -o nounset
#数据库的端口
PORTS=( 3306 3308 )
PSW="123456"
REPL_USER="repl"
REPL_USER_PSW="repl123456"
#ANSIBLE_HOST_NAMES=( "168" ) #同 SLAVE_HOST_IP ，需要在/etc/ansible/hosts里配置
ANSIBLE_HOST_NAMES=( "159" ) #同 SLAVE_HOST_IP ，需要在/etc/ansible/hosts里配置
#[root@t168 ~]# more /etc/ansible/hosts
#[159]
#192.168.1.159
##############################################
# #
# 配置修改开始 #
# #
##############################################
LOCAL_HOST_IP="192.168.1.168" ##升级为主库的现有IP地址
MASTER_HOST_IP='192.168.1.170' ## 主库的IP地址
SLAVE_HOST_IP="192.168.1.157" ##原另一个从库的IP
MYSQL_BIN=`which mysql`
MYSQL_MASTER_BIN='${MYSQL_BIN} -uroot -p${PSW}'
##############################################
# #
# 配置修改结束 #
# #
##############################################
# define restricted path
PATH="/usr/local/mysql/bin:/bin:/usr/bin:/sbin:/usr/sbin"
# adirname - return absolute dirname of given file
adirname() { odir=`pwd`; cd `dirname $1`; pwd; cd "${odir}"; }
MYNAM=`basename "$0"`
MYDIR=`adirname "$0"`
MYLOG_PATH="${MYDIR}/logs"
MYLOG="${MYLOG_PATH}/${MYNAM}_`date +%F`.log"
for D in ${MYLOG_PATH}
do
if [ ! -d ${D} ] ; then
mkdir -p ${D}
echo -e "Mkdir ${D}" >> ${MYLOG}
fi
done
# ---------
# functions
# ---------
#日志函数
function L(){
message="$(date -d today +"%Y-%m-%d %H:%M:%S") - $1"
echo -e "\033[34m $message \033[0m" && echo $message >> ${MYLOG}
}
#主库挂，从库升级为主库
#1、keepalived 切vip到从库上
#2、脚本停止slave同步
#3、重看当前从库的pos和bin-logs记录
#4、修改从库为读写状态
#5、修改另一个从库来连接新的主库并查看同步状态
function CheckMasterIP(){
#1、keepalived 切vip到主库上。keepalived为非抢占模式，切回来不自动切回去。
#获取本地是否有vip
VIP=`/bin/hostname -I|grep "${MASTER_HOST_IP}"`
if [ "x${VIP}" == "x" ]; then
L "vip${MASTER_HOST_IP}没有切到这个服务器上。"
exit 1
fi
L "vip${MASTER_HOST_IP}切换到该服务器上成功"
return 0
}
function StopSlave(){
#2、停止当前slave同步
${MYSQL_MASTER_BIN} -e "stop slave"
${MYSQL_MASTER_BIN} -e "reset slave"
L "${MYSQL_MASTER_BIN} -e \"stop slave\""
SLAVESTATUS=`${MYSQL_MASTER_BIN} -e "show slave status\G"|grep "Slave_SQL_Running"|grep -v "State"|awk '{print $NF}'`
L "${SLAVESTATUS} ${MYSQL_MASTER_BIN} -e \"show slave status\G\"|grep \"Slave_SQL_Running\"|grep -v \"State\"|awk '{print \$NF}'"
if [ "${SLAVESTATUS}" != "No" ];then
L "当前db${IPORT}停止slave失败。"
exit 1
fi
L "停止当前DB${IPORT}的slave同步成功"
return 0
}
function GetNewMasterPosNum(){
#3、重看当前升级为主库的posnum记录
#posnum
Pos=`${MYSQL_MASTER_BIN} -e "show master status\G"|grep "Position"|awk -F " " '{print $2}'`
L "新的master库${IPORT}的pos是${Pos}"
return ${Pos}
}
function AlterReadOnlyStatus(){
#4、修改当前库为读写状态
${MYSQL_MASTER_BIN} -e "set global read_only=0"
#修改网卡配置文件，但不重启（防止该服务器重启后临时IP地址丢失）
sed -i "s/read_only = 1/read_only = 0/g" /etc/my${IPORT}.cnf
L "sed -i \"s/read_only = 1/read_only = 0/g\" /etc/my${IPORT}.cnf"
ReadOnlyStatus=`${MYSQL_MASTER_BIN} -e "show variables like \"read_only\""|awk '{print $NF}'|sed -n '$p'`
ReadOnlyConfig=`grep "read_only" /etc/my${IPORT}.cnf |awk '{print $NF}'`
#echo "${ReadOnlyStatus}"
if [ "${ReadOnlyStatus}" != "OFF" ] && [ "${ReadOnlyConfig}" != "0" ];then
L "修改当前数据库${IPORT}的读写状态失败。"
exit 1
fi
L "修改当前数据库${IPORT}的读写状态成功。"
return 0
}
function AnsibleSlaveConnect(){
#5、修改另一个从库来连接新的主库并查看同步状态
for ANSIBLE_HOST_NAME in ${ANSIBLE_HOST_NAMES[@]}
do
ansible ${ANSIBLE_HOST_NAME} -m shell -a "${MYSQL_MASTER_BIN} -e \"stop slave\""
L "ansible ${ANSIBLE_HOST_NAME} -m shell -a \"${MYSQL_MASTER_BIN} -e \"stop slave\"\""
L "另一个从库${IPORT}停止原主从同步成功"
ansible ${ANSIBLE_HOST_NAME} -m shell -a "${MYSQL_MASTER_BIN} -e \"reset slave\""
L "另一个从库${IPORT}重置原主从同步成功"
ansible ${ANSIBLE_HOST_NAME} -m shell -a "${MYSQL_MASTER_BIN} -e \"set global read_only=1\""
L "获取新主库${IPORT}的bin-logs记录"
LogFile=`${MYSQL_MASTER_BIN} -e "show master status\G"|egrep "File"|awk -F " " '{print $2}'`
L "获取新主库${IPORT}的Pos记录"
GetNewMasterPosNum
ansible ${ANSIBLE_HOST_NAME} -m shell -a "${MYSQL_MASTER_BIN} -e \"change master to master_host='${MASTER_HOST_IP}',master_port=${IPORT},master_user='${REPL_USER}',master_password='${REPL_USER_PSW}',MASTER_LOG_FILE='${LogFile}',MASTER_LOG_POS=${Pos}\""
L "另一个从库${IPORT}指定新的主库"
L "ansible ${ANSIBLE_HOST_NAME} -m shell -a \"${MYSQL_MASTER_BIN} -e \"change master to master_host='${MASTER_HOST_IP}',master_port=${IPORT},master_user='${REPL_USER}',master_password='${REPL_USER_PSW}',MASTER_LOG_FILE=${LogFile},MASTER_LOG_POS=${Pos}\"\""
ansible ${ANSIBLE_HOST_NAME} -m shell -a "${MYSQL_MASTER_BIN} -e \"start slave\""
L "另一个从库${IPORT}开启同步"
SlaveRsyncIp=`ansible ${ANSIBLE_HOST_NAME} -m shell -a "${MYSQL_MASTER_BIN} -e \"show slave status\G\""|grep "Master_Host"|awk '{print $NF}'`
L "ansible ${ANSIBLE_HOST_NAME} -m shell -a "${MYSQL_MASTER_BIN} -e \"show slave status\G\""|grep "Master_Host"|awk '{print \$NF}'"
L "SlaveRsyncIp ${SlaveRsyncIp} ,${MASTER_HOST_IP}"
if [ "${SlaveRsyncIp}" != "${MASTER_HOST_IP}" ];then
L "同步主库${IPORT}的IP错误"
exit 1
fi
L "另一个从库${IPORT}配置完成，并开始同步"
return 0
done
}
##############################################
# #
# 始设置本机为master db #
# #
##############################################
function MasterDB(){
for IPORT in ${PORTS[@]}
do
MYSQL_MASTER_BIN="`which mysql` -uroot -p${PSW} -P${IPORT} -S /tmp/mysql${IPORT}.sock"
CheckMasterIP
if [ $? != 0 ];then
L "增加临时IP或修改IP配置文件失败."
exit 1
fi
StopSlave
if [ $? != 0 ];then
L "停止${IPORT}的slave同步失败."
exit 1
fi
AlterReadOnlyStatus
if [ $? != 0 ];then
L "设置master${IPORT}库为读写库失败."
exit 1
fi
AnsibleSlaveConnect
if [ $? != 0 ];then
L "修改另一从库${IPORT}的master库为新的master库失败."
exit 1
fi
L "${IPORT}设置为新的主库完成！"
done
}
##############################################
# #
# 开始执行脚本 #
# #
##############################################
MasterDB
至此，实现了mysql vip的自动切换，同时其他从库从新的master库上同步
