论坛 › MySQL运维 › mysql5.6基于GTID模式之高可用架构搭建-MHA（mha0.56）

ewreew 发表于 2017-3-15 10:44:27

mysql5.6基于GTID模式之高可用架构搭建-MHA（mha0.56）

一、测试环境部署：
mysql1:192.168.110.131   作为master

mysql2:192.168.110.132   作为slave

mysql3:192.168.110.130   作为slave，同时作为MHA的管理机
虚拟IP：192.168.110.100
二、mysql主从环境搭建和MHA安装
1、mysql主从搭建自行搭建（基于GTID复制，打开log_bin,复制规则默认，复制所有库表），这里不再说明。
2、安装MHA节点软件：rpm -ivh mha4mysql-node-0.56-0.el6.noarch.rpm
3、安装MHA管理节点：rpm -ivh mha4mysql-manager-0.56-0.el6.noarch.rpm
    yum install perl-DBD-MySQL perl-Config-Tiny perl-Log-Dispatch perl-Parallel-ForkManager perl-Time-HiRes -y
4、配置MHA管理目录和配置：
mkdir -p /etc/masterha   #工作目录
mkdir -p /var/log/masterha/app1/ #日志目录

cat /etc/masterha/app1.cnf #配置文件

manager_log=/var/log/masterha/app1/manager.log
manager_workdir=/var/log/masterha/app1
master_ip_failover_script="/usr/bin/master_ip_failover"
ping_interval=1
remote_workdir=/tmp
repl_password=repl
repl_user=repl
secondary_check_script="/usr/bin/masterha_secondary_check -s 192.168.110.132 -s 192.168.110.130 --user=root --master_host=mysql1 --master_ip=192.168.110.131 --master_port=3306"
ssh_user=root
user=mha
password=mha


hostname=mysql1
master_binlog_dir="/data/mysql/mysql3306/binlog"
port=3306


hostname=mysql2
master_binlog_dir="/data/mysql/mysql3306/binlog"
port=3306
candidate_master=1
check_repl_delay=0


hostname=ansible
master_binlog_dir="/data/mysql/mysql3306/binlog"
port=3306


hostname=mysql2
master_binlog_dir=/tmp/binlog
ignore_fail=1
no_master=1

5、配置MHA管理数据库账号mha:
mysql>GRANT ALL PRIVILEGES ON *.* TO 'mha'@'192.168.110.%' IDENTIFIED BY 'mha';
6、配置主库的binlog实时备份，脚本部署在mysql2上（非必须，如无，请去掉app1.cnf里面的配置信息）：

cat binlog.sh
#!/bin/sh
BACKUP_BIN="/usr/bin/mysqlbinlog"
LOCAL_BACKUP_DIR="/tmp/binlog"
BACKUP_LOG="/tmp/binlog/binary.log"

REMOTE_HOST="192.168.110.131"
REMOTE_PORT="3306"
REMOTE_USER="root"
REMOTE_PASS="root"
FIRST_BINLOG="mysql-bin.000001"

#time to wait before reconnecting after failure
SLEEP_SECONDS=10

##create local_backup_dir if necessary
mkdir -p ${LOCAL_BACKUP_DIR}
cd ${LOCAL_BACKUP_DIR}

## 运行while循环，连接断开后等待指定时间，重新连接
while :
do
if [ `ls -A "${LOCAL_BACKUP_DIR}" |wc -l` -eq 0 ];then
   LAST_FILE=${FIRST_BINLOG}
else
   LAST_FILE=`ls -l ${LOCAL_BACKUP_DIR} | grep -v backuplog | grep mysql-bin |tail -n 1 |awk '{print $9}'`
fi
${BACKUP_BIN} --raw --read-from-remote-server --stop-never --host=${REMOTE_HOST} --port=${REMOTE_PORT} --user=${REMOTE_USER} --password=${REMOTE_PASS} ${LAST_FILE}

echo "`date +"%Y/%m/%d %H:%M:%S"` mysqlbinlog停止，返回代码：$?" | tee -a ${BACKUP_LOG}
echo "${SLEEP_SECONDS}秒后再次连接并继续备份" | tee -a ${BACKUP_LOG}
sleep ${SLEEP_SECONDS}
done
7、修改/usr/bin/master_ip_failover脚本，这里采用脚本切换VIP的方式实现VIP漂移。

#!/usr/bin/env perl

use strict;
use warnings FATAL => 'all';

use Getopt::Long;

my (
    $command,          $ssh_user,      $orig_master_host, $orig_master_ip,
    $orig_master_port, $new_master_host, $new_master_ip,    $new_master_port
);

my $gateway = '192.168.110.2';
my $vip = '192.168.110.100';# Virtual IP
my $key = "2";
my $ssh_start_vip = "/sbin/ifconfig eth0:$key $vip;/sbin/arping -I eth0 -c 3 -s $vip $gateway >/dev/null 2>&1";
my $ssh_stop_vip = "/sbin/ifconfig eth0:$key down";
$ssh_user = "root";

GetOptions(
    'command=s'          => \$command,
    'ssh_user=s'         => \$ssh_user,
    'orig_master_host=s' => \$orig_master_host,
    'orig_master_ip=s'   => \$orig_master_ip,
    'orig_master_port=i' => \$orig_master_port,
    'new_master_host=s'=> \$new_master_host,
    'new_master_ip=s'    => \$new_master_ip,
    'new_master_port=i'=> \$new_master_port,
);

exit &main();

sub main {

    print "\n\nIN SCRIPT TEST====$ssh_stop_vip==$ssh_start_vip===\n\n";

    if ( $command eq "stop" || $command eq "stopssh" ) {

      # $orig_master_host, $orig_master_ip, $orig_master_port are passed.
      # If you manage master ip address at global catalog database,
      # invalidate orig_master_ip here.
      my $exit_code = 1;

      #eval {
      #    print "Disabling the VIP on old master: $orig_master_host \n";
      #    &stop_vip();
      #    $exit_code = 0;
      #};


      eval {
                print "Disabling the VIP on old master: $orig_master_host \n";
                #my $ping=`ping -c 1 10.0.0.13 | grep "packet loss" | awk -F',' '{print $3}' | awk '{print $1}'`;
                #if ( $ping le "90.0%" && $ping gt "0.0%" ){
                #$exit_code = 0;
                #}
                #else {

                &stop_vip();

                # updating global catalog, etc
                $exit_code = 0;

                #}
      };


      if ($@) {
            warn "Got Error: $@\n";
            exit $exit_code;
      }
      exit $exit_code;
    }
    elsif ( $command eq "start" ) {

      # all arguments are passed.
      # If you manage master ip address at global catalog database,
      # activate new_master_ip here.
      # You can also grant write access (create user, set read_only=0, etc) here.
      my $exit_code = 10;
      eval {
            print "Enabling the VIP - $vip on the new master - $new_master_host \n";
            &start_vip();
            $exit_code = 0;
      };
      if ($@) {
            warn $@;
            exit $exit_code;
      }
      exit $exit_code;
    }
    elsif ( $command eq "status" ) {
      print "Checking the Status of the script.. OK \n";
      `ssh $ssh_user\@$orig_master_ip \" $ssh_start_vip \"`;
      exit 0;
    }
    else {
      &usage();
      exit 1;
    }
}

# A simple system call that enable the VIP on the new master
sub start_vip() {
    `ssh $ssh_user\@$new_master_host \" $ssh_start_vip \"`;
}

# A simple system call that disable the VIP on the old_master
sub stop_vip() {
    `ssh $ssh_user\@$orig_master_host \" $ssh_stop_vip \"`;
}

sub usage {
    print
    "Usage: master_ip_failover --command=start|stop|stopssh|status --orig_master_host=host --orig_master_ip=ip --orig_master_port=port --new_master_host=host --new_master_ip=ip --new_master_port=port\n";
}

# the end.

8、配置SSH免密登陆：(以ansible为例)
ssh-keygen -t rsa
ssh-copy-id -i ~/.ssh/id_rsa.pub root@mysql1
ssh-copy-id -i ~/.ssh/id_rsa.pub root@mysql2
ssh-copy-id -i ~/.ssh/id_rsa.pub root@ansible
基本上MHA manager，MNA node，以及二次检测的节点，都需要互相信任
注：这里配置免密的时候采用主机名，那在app1.cnf里的hostname对应也要配置成主机名。如果是IP，那就配置成IP。

9、配置sysbench压测工具，模拟生产对数据库进行压测。
安装sysbench请自行百度，这里只提供压测脚本。
mysql>grant all privileges on sbtest1.* on root@'192.168.110.%' identified by 'root';

vim sysbench.sh

#!/bin/sh


function prepare(){
sysbench --test=oltp --mysql-host=192.168.110.100 --mysql-port=3306 --mysql-user=root --mysql-password=root --mysql-db=sbtest1 --oltp-num-tables=10 --oltp-table-size=500000 --report-interval=10--max-requests=0 --oltp-test-mode=nontrx --oltp-nontrx-mode=select --oltp-read-only=off --max-time=30 --num-threads=10 prepare
}

function run(){
sysbench --test=oltp --mysql-host=192.168.110.100 --mysql-port=3306 --mysql-user=root --mysql-password=root --mysql-db=sbtest1 --oltp-num-tables=10 --oltp-table-size=500000 --report-interval=10--max-requests=0 --oltp-test-mode=nontrx --oltp-nontrx-mode=select --oltp-read-only=off --max-time=30 --num-threads=10 run
}

function cleanup(){
sysbench --test=oltp --mysql-host=192.168.110.100 --mysql-port=3306 --mysql-user=root --mysql-password=root --mysql-db=sbtest1 --oltp-num-tables=10 --oltp-table-size=500000 --report-interval=10--max-requests=0 --oltp-test-mode=nontrx --oltp-nontrx-mode=select --oltp-read-only=off --max-time=30 --num-threads=10 cleanup
}

case "$1" in
prepare)
prepare
;;
run)
run
;;
cleanup)
cleanup
;;
*)
echo "Usage: $0 prepare|run|cleanup"
esac

脚本对VIP所在的mysql即主库进行压测。

10、检测SSH状态：

11、检测复制状态：
# masterha_check_repl --conf=/etc/masterha/app1.cnf
Tue Mar 14 17:29:38 2017 - Global configuration file /etc/masterha_default.cnf not found. Skipping.
Tue Mar 14 17:29:38 2017 - Reading application default configuration from /etc/masterha/app1.cnf..
Tue Mar 14 17:29:38 2017 - Reading server configuration from /etc/masterha/app1.cnf..
Tue Mar 14 17:29:38 2017 - MHA::MasterMonitor version 0.56.
Tue Mar 14 17:29:38 2017 - GTID failover mode = 0
Tue Mar 14 17:29:38 2017 - Dead Servers:
Tue Mar 14 17:29:38 2017 - Alive Servers:
Tue Mar 14 17:29:38 2017 -    mysql1(192.168.110.131:3306)
Tue Mar 14 17:29:38 2017 -    mysql2(192.168.110.132:3306)
Tue Mar 14 17:29:38 2017 -    ansible(192.168.110.130:3306)
Tue Mar 14 17:29:38 2017 - Alive Slaves:
Tue Mar 14 17:29:38 2017 -    mysql2(192.168.110.132:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 17:29:38 2017 -    GTID ON
Tue Mar 14 17:29:38 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 17:29:38 2017 -    Primary candidate for the new Master (candidate_master is set)
Tue Mar 14 17:29:38 2017 -    ansible(192.168.110.130:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 17:29:38 2017 -    GTID ON
Tue Mar 14 17:29:38 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 17:29:38 2017 - Current Alive Master: mysql1(192.168.110.131:3306)
Tue Mar 14 17:29:38 2017 - Checking slave configurations..
Tue Mar 14 17:29:38 2017 - read_only=1 is not set on slave mysql2(192.168.110.132:3306).
Tue Mar 14 17:29:38 2017 - read_only=1 is not set on slave ansible(192.168.110.130:3306).
Tue Mar 14 17:29:38 2017 - Checking replication filtering settings..
Tue Mar 14 17:29:38 2017 - binlog_do_db= , binlog_ignore_db=
Tue Mar 14 17:29:38 2017 - Replication filtering check ok.
Tue Mar 14 17:29:38 2017 - GTID (with auto-pos) is not supported
Tue Mar 14 17:29:38 2017 - Starting SSH connection tests..
Tue Mar 14 17:29:40 2017 - All SSH connection tests passed successfully.
Tue Mar 14 17:29:40 2017 - Checking MHA Node version..
Tue Mar 14 17:29:40 2017 - Version check ok.
Tue Mar 14 17:29:40 2017 - Checking SSH publickey authentication settings on the current master..
Tue Mar 14 17:29:40 2017 - HealthCheck: SSH to mysql1 is reachable.
Tue Mar 14 17:29:40 2017 - Checking recovery script configurations on mysql1(192.168.110.131:3306)..
Tue Mar 14 17:29:40 2017 -    Executing command: save_binary_logs --command=test --start_pos=4 --binlog_dir=/data/mysql/mysql3306/binlog --output_file=/tmp/save_binary_logs_test --manager_version=0.56 --start_file=mysql-bin.000001
Tue Mar 14 17:29:40 2017 -    Connecting to root@192.168.110.131(mysql1:22)..
Creating /tmp if not exists..    ok.
Checking output directory is accessible or not..
   ok.
Binlog found at /data/mysql/mysql3306/binlog, up to mysql-bin.000001
Tue Mar 14 17:29:41 2017 - Binlog setting check done.
Tue Mar 14 17:29:41 2017 - Checking SSH publickey authentication and checking recovery script configurations on all alive slave servers..
Tue Mar 14 17:29:41 2017 -    Executing command : apply_diff_relay_logs --command=test --slave_user='mha' --slave_host=mysql2 --slave_ip=192.168.110.132 --slave_port=3306 --workdir=/tmp --target_version=5.6.34-log --manager_version=0.56 --relay_dir=/data/mysql/mysql3306/relaylog --current_relay_log=slave-relay-bin.000002--slave_pass=xxx
Tue Mar 14 17:29:41 2017 -    Connecting to root@192.168.110.132(mysql2:22)..
Checking slave recovery environment settings..
    Relay log found at /data/mysql/mysql3306/relaylog, up to slave-relay-bin.000002
    Temporary relay log file is /data/mysql/mysql3306/relaylog/slave-relay-bin.000002
    Testing mysql connection and privileges..Warning: Using a password on the command line interface can be insecure.
done.
    Testing mysqlbinlog output.. done.
    Cleaning up test file(s).. done.
Tue Mar 14 17:29:42 2017 -    Executing command : apply_diff_relay_logs --command=test --slave_user='mha' --slave_host=ansible --slave_ip=192.168.110.130 --slave_port=3306 --workdir=/tmp --target_version=5.6.34-log --manager_version=0.56 --relay_dir=/data/mysql/mysql3306/relaylog --current_relay_log=slave-relay-bin.000002--slave_pass=xxx
Tue Mar 14 17:29:42 2017 -    Connecting to root@192.168.110.130(ansible:22)..
Checking slave recovery environment settings..
    Relay log found at /data/mysql/mysql3306/relaylog, up to slave-relay-bin.000002
    Temporary relay log file is /data/mysql/mysql3306/relaylog/slave-relay-bin.000002
    Testing mysql connection and privileges..Warning: Using a password on the command line interface can be insecure.
done.
    Testing mysqlbinlog output.. done.
    Cleaning up test file(s).. done.
Tue Mar 14 17:29:42 2017 - Slaves settings check done.
Tue Mar 14 17:29:42 2017 -
mysql1(192.168.110.131:3306) (current master)
+--mysql2(192.168.110.132:3306)
+--ansible(192.168.110.130:3306)

Tue Mar 14 17:29:42 2017 - Checking replication health on mysql2..
Tue Mar 14 17:29:42 2017 - ok.
Tue Mar 14 17:29:42 2017 - Checking replication health on ansible..
Tue Mar 14 17:29:42 2017 - ok.
Tue Mar 14 17:29:42 2017 - Checking master_ip_failover_script status:
Tue Mar 14 17:29:42 2017 -    /usr/bin/master_ip_failover --command=status --ssh_user=root --orig_master_host=mysql1 --orig_master_ip=192.168.110.131 --orig_master_port=3306


IN SCRIPT TEST====/sbin/ifconfig eth0:2 down==/sbin/ifconfig eth0:2 192.168.110.100;/sbin/arping -I eth0 -c 3 -s 192.168.110.100 192.168.110.2 >/dev/null 2>&1===

Checking the Status of the script.. OK
Tue Mar 14 17:29:45 2017 - OK.
Tue Mar 14 17:29:45 2017 - shutdown_script is not defined.
Tue Mar 14 17:29:45 2017 - Got exit code 0 (Not master dead).

MySQL Replication Health is OK.

13、启动、关闭和检测manager状态：
vim masterha_manager.sh

#!/bin/sh

function start(){
nohup /usr/bin/masterha_manager --conf=/etc/masterha/app1.cnf --ignore_last_failover < /dev/null > /var/log/masterha/app1/manager.log 2>&1 &
}

function stop(){
/usr/bin/masterha_stop --conf=/etc/masterha/app1.cnf
}

function status(){
/usr/bin/masterha_check_status --conf=/etc/masterha/app1.cnf
}

case "$1" in
start)
start
;;
stop)
stop
;;
status)
status
;;
*)
echo "Usage: $0 start|stop"
esac


14、初始化VIP：
首次启用MHA，需要提前手工配置VIP在mysql主库上。
shell>/sbin/ifconfig eth0:2 192.168.110.100

15、场景测试：

[*]启动manager进程：sh masterha_manager.sh start


[*]启动压测脚本：sh sysbench.sh prepare

[*]中断mysql2的slave线程,以造成mysql2有大量延迟：mysql>stop slave;

[*]压测一段时间

[*]启动mysql2的slave线程，检查slave状态。并可以强行杀掉主库进程（pkill -9 mysql），即使sysbench.sh还未跑完的时候

[*]观察此时MHA manager进程状态和日志信息。

[*]检验VIP是否切换，检查主库是否切换，从库是否从新主库那里复制。


注：即使在配置文件中配置mysql2为candidate_master=1和check_repl_delay=0，但是mysql2延迟量太大的话，MHA还是会选择拥有最新数据的slave作为master。

日志如下：
# cat manager.log
Tue Mar 14 16:25:27 2017 - Global configuration file /etc/masterha_default.cnf not found. Skipping.
Tue Mar 14 16:25:27 2017 - Reading application default configuration from /etc/masterha/app1.cnf..
Tue Mar 14 16:25:27 2017 - Reading server configuration from /etc/masterha/app1.cnf..
Tue Mar 14 16:26:28 2017 - Global configuration file /etc/masterha_default.cnf not found. Skipping.
Tue Mar 14 16:26:28 2017 - Reading application default configuration from /etc/masterha/app1.cnf..
Tue Mar 14 16:26:28 2017 - Reading server configuration from /etc/masterha/app1.cnf..
ive Servers:
Tue Mar 14 16:25:28 2017 -    mysql1(192.168.110.131:3306)
Tue Mar 14 16:25:28 2017 -    mysql2(192.168.110.132:3306)
Tue Mar 14 16:25:28 2017 -    ansible(192.168.110.130:3306)
Tue Mar 14 16:25:28 2017 - Alive Slaves:
Tue Mar 14 16:25:28 2017 -    mysql2(192.168.110.132:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 16:25:28 2017 -    GTID ON
Tue Mar 14 16:25:28 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 16:25:28 2017 -    Primary candidate for the new Master (candidate_master is set)
Tue Mar 14 16:25:28 2017 -    ansible(192.168.110.130:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 16:25:28 2017 -    GTID ON
Tue Mar 14 16:25:28 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 16:25:28 2017 - Current Alive Master: mysql1(192.168.110.131:3306)
Tue Mar 14 16:25:28 2017 - Checking slave configurations..
Tue Mar 14 16:25:28 2017 - read_only=1 is not set on slave mysql2(192.168.110.132:3306).
Tue Mar 14 16:25:28 2017 - read_only=1 is not set on slave ansible(192.168.110.130:3306).
Tue Mar 14 16:25:28 2017 - Checking replication filtering settings..
Tue Mar 14 16:25:28 2017 - binlog_do_db= , binlog_ignore_db=
Tue Mar 14 16:25:28 2017 - Replication filtering check ok.
Tue Mar 14 16:25:28 2017 - GTID (with auto-pos) is supported. Skipping all SSH and Node package checking.
Tue Mar 14 16:25:29 2017 - HealthCheck: SSH to mysql2 is reachable.
Tue Mar 14 16:25:29 2017 - Binlog server mysql2 is reachable.
Tue Mar 14 16:25:29 2017 - Checking recovery script configurations on mysql2(192.168.110.132:3306)..
Tue Mar 14 16:25:29 2017 -    Executing command: save_binary_logs --command=test --start_pos=4 --binlog_dir=/tmp/binlog --output_file=/tmp/save_binary_logs_test --manager_version=0.56 --start_file=mysql-bin.000001
Tue Mar 14 16:25:29 2017 -    Connecting to root@192.168.110.132(mysql2:22)..
Creating /tmp if not exists..    ok.
Checking output directory is accessible or not..
   ok.
Binlog found at /tmp/binlog, up to mysql-bin.000015
Tue Mar 14 16:25:29 2017 - Binlog setting check done.
Tue Mar 14 16:25:29 2017 - Checking SSH publickey authentication settings on the current master..
Tue Mar 14 16:25:34 2017 - HealthCheck: Got timeout on checking SSH connection to mysql1! at /usr/share/perl5/vendor_perl/MHA/HealthCheck.pm line 342.
Tue Mar 14 16:25:34 2017 -
mysql1(192.168.110.131:3306) (current master)
+--mysql2(192.168.110.132:3306)
+--ansible(192.168.110.130:3306)

Tue Mar 14 16:25:34 2017 - Checking master_ip_failover_script status:
Tue Mar 14 16:25:34 2017 -    /usr/bin/master_ip_failover --command=status --ssh_user=root --orig_master_host=mysql1 --orig_master_ip=192.168.110.131 --orig_master_port=3306


IN SCRIPT TEST====/sbin/ifconfig eth0:2 down==/sbin/ifconfig eth0:2 192.168.110.100;/sbin/arping -I eth0 -c 3 -s 192.168.110.100 192.168.110.2 >/dev/null 2>&1===

Checking the Status of the script.. OK
Tue Mar 14 16:25:49 2017 - OK.
Tue Mar 14 16:25:49 2017 - shutdown_script is not defined.
Tue Mar 14 16:25:49 2017 - Set master ping interval 1 seconds.
Tue Mar 14 16:25:49 2017 - Set secondary check script: /usr/bin/masterha_secondary_check -s 192.168.110.132 -s 192.168.110.130 --user=root --master_host=mysql1 --master_ip=192.168.110.131 --master_port=3306
Tue Mar 14 16:25:49 2017 - Starting ping health check on mysql1(192.168.110.131:3306)..
Tue Mar 14 16:25:49 2017 - Ping(SELECT) succeeded, waiting until MySQL doesn't respond..
Tue Mar 14 16:26:25 2017 - Got error on MySQL select ping: 2006 (MySQL server has gone away)
Tue Mar 14 16:26:25 2017 - Executing SSH check script: exit 0
Tue Mar 14 16:26:25 2017 - Executing secondary network check script: /usr/bin/masterha_secondary_check -s 192.168.110.132 -s 192.168.110.130 --user=root --master_host=mysql1 --master_ip=192.168.110.131 --master_port=3306--user=root--master_host=mysql1--master_ip=192.168.110.131--master_port=3306 --master_user=mha --master_password=mha --ping_type=SELECT
Monitoring server 192.168.110.132 is reachable, Master is not reachable from 192.168.110.132. OK.
Tue Mar 14 16:26:26 2017 - HealthCheck: SSH to mysql1 is reachable.
Monitoring server 192.168.110.130 is reachable, Master is not reachable from 192.168.110.130. OK.
Tue Mar 14 16:26:26 2017 - Master is not reachable from all other monitoring servers. Failover should start.
Tue Mar 14 16:26:26 2017 - Got error on MySQL connect: 2013 (Lost connection to MySQL server at 'reading initial communication packet', system error: 111)
Tue Mar 14 16:26:26 2017 - Connection failed 2 time(s)..
Tue Mar 14 16:26:27 2017 - Got error on MySQL connect: 2013 (Lost connection to MySQL server at 'reading initial communication packet', system error: 111)
Tue Mar 14 16:26:27 2017 - Connection failed 3 time(s)..
Tue Mar 14 16:26:28 2017 - Got error on MySQL connect: 2013 (Lost connection to MySQL server at 'reading initial communication packet', system error: 111)
Tue Mar 14 16:26:28 2017 - Connection failed 4 time(s)..
Tue Mar 14 16:26:28 2017 - Master is not reachable from health checker!
Tue Mar 14 16:26:28 2017 - Master mysql1(192.168.110.131:3306) is not reachable!
Tue Mar 14 16:26:28 2017 - SSH is reachable.
Tue Mar 14 16:26:28 2017 - Connecting to a master server failed. Reading configuration file /etc/masterha_default.cnf and /etc/masterha/app1.cnf again, and trying to connect to all servers to check server status..
Tue Mar 14 16:26:28 2017 - Global configuration file /etc/masterha_default.cnf not found. Skipping.
Tue Mar 14 16:26:28 2017 - Reading application default configuration from /etc/masterha/app1.cnf..
Tue Mar 14 16:26:28 2017 - Reading server configuration from /etc/masterha/app1.cnf..
Tue Mar 14 16:26:28 2017 - GTID failover mode = 1
Tue Mar 14 16:26:28 2017 - Dead Servers:
Tue Mar 14 16:26:28 2017 -    mysql1(192.168.110.131:3306)
Tue Mar 14 16:26:28 2017 - Alive Servers:
Tue Mar 14 16:26:28 2017 -    mysql2(192.168.110.132:3306)
Tue Mar 14 16:26:28 2017 -    ansible(192.168.110.130:3306)
Tue Mar 14 16:26:28 2017 - Alive Slaves:
Tue Mar 14 16:26:28 2017 -    mysql2(192.168.110.132:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 16:26:28 2017 -    GTID ON
Tue Mar 14 16:26:28 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 16:26:28 2017 -    Primary candidate for the new Master (candidate_master is set)
Tue Mar 14 16:26:28 2017 -    ansible(192.168.110.130:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 16:26:28 2017 -    GTID ON
Tue Mar 14 16:26:28 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 16:26:28 2017 - Checking slave configurations..
Tue Mar 14 16:26:28 2017 - read_only=1 is not set on slave mysql2(192.168.110.132:3306).
Tue Mar 14 16:26:28 2017 - read_only=1 is not set on slave ansible(192.168.110.130:3306).
Tue Mar 14 16:26:28 2017 - Checking replication filtering settings..
Tue Mar 14 16:26:28 2017 - Replication filtering check ok.
Tue Mar 14 16:26:28 2017 - Master is down!
Tue Mar 14 16:26:28 2017 - Terminating monitoring script.
Tue Mar 14 16:26:28 2017 - Got exit code 20 (Master dead).
Tue Mar 14 16:26:28 2017 - MHA::MasterFailover version 0.56.
Tue Mar 14 16:26:28 2017 - Starting master failover.
Tue Mar 14 16:26:28 2017 -
Tue Mar 14 16:26:28 2017 - * Phase 1: Configuration Check Phase..
Tue Mar 14 16:26:28 2017 -
Tue Mar 14 16:26:28 2017 - HealthCheck: SSH to mysql2 is reachable.
Tue Mar 14 16:26:28 2017 - Binlog server mysql2 is reachable.
Tue Mar 14 16:26:29 2017 - GTID failover mode = 1
Tue Mar 14 16:26:29 2017 - Dead Servers:
Tue Mar 14 16:26:29 2017 -    mysql1(192.168.110.131:3306)
Tue Mar 14 16:26:29 2017 - Checking master reachability via MySQL(double check)...
Tue Mar 14 16:26:29 2017 - ok.
Tue Mar 14 16:26:29 2017 - Alive Servers:
Tue Mar 14 16:26:29 2017 -    mysql2(192.168.110.132:3306)
Tue Mar 14 16:26:29 2017 -    ansible(192.168.110.130:3306)
Tue Mar 14 16:26:29 2017 - Alive Slaves:
Tue Mar 14 16:26:29 2017 -    mysql2(192.168.110.132:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 16:26:29 2017 -    GTID ON
Tue Mar 14 16:26:29 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 16:26:29 2017 -    Primary candidate for the new Master (candidate_master is set)
Tue Mar 14 16:26:29 2017 -    ansible(192.168.110.130:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 16:26:29 2017 -    GTID ON
Tue Mar 14 16:26:29 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 16:26:29 2017 - Starting GTID based failover.
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - ** Phase 1: Configuration Check Phase completed.
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - * Phase 2: Dead Master Shutdown Phase..
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - Forcing shutdown so that applications never connect to the current master..
Tue Mar 14 16:26:29 2017 - Executing master IP deactivation script:
Tue Mar 14 16:26:29 2017 -    /usr/bin/master_ip_failover --orig_master_host=mysql1 --orig_master_ip=192.168.110.131 --orig_master_port=3306 --command=stopssh --ssh_user=root


IN SCRIPT TEST====/sbin/ifconfig eth0:2 down==/sbin/ifconfig eth0:2 192.168.110.100;/sbin/arping -I eth0 -c 3 -s 192.168.110.100 192.168.110.2 >/dev/null 2>&1===

Disabling the VIP on old master: mysql1
Tue Mar 14 16:26:29 2017 - done.
Tue Mar 14 16:26:29 2017 - shutdown_script is not set. Skipping explicit shutting down of the dead master.
Tue Mar 14 16:26:29 2017 - * Phase 2: Dead Master Shutdown Phase completed.
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - * Phase 3: Master Recovery Phase..
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - * Phase 3.1: Getting Latest Slaves Phase..
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - The latest binary log file/position on all slaves is mysql-bin.000001:133309374
Tue Mar 14 16:26:29 2017 - Retrieved Gtid Set: c897448e-07c8-11e7-8af3-000c29918f29:1-244
Tue Mar 14 16:26:29 2017 - Latest slaves (Slaves that received relay log files to the latest):
Tue Mar 14 16:26:29 2017 -    ansible(192.168.110.130:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 16:26:29 2017 -    GTID ON
Tue Mar 14 16:26:29 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 16:26:29 2017 - The oldest binary log file/position on all slaves is mysql-bin.000001:31201438
Tue Mar 14 16:26:29 2017 - Retrieved Gtid Set: c897448e-07c8-11e7-8af3-000c29918f29:1-81
Tue Mar 14 16:26:29 2017 - Oldest slaves:
Tue Mar 14 16:26:29 2017 -    mysql2(192.168.110.132:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 16:26:29 2017 -    GTID ON
Tue Mar 14 16:26:29 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 16:26:29 2017 -    Primary candidate for the new Master (candidate_master is set)
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - * Phase 3.3: Determining New Master Phase..
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - Slave mysql2(192.168.110.132:3306) SQL Thread delays too much. Latest log file:mysql-bin.000001:133309374, Current log file:mysql-bin.000001:31152313. This server is not selected as a new master because recovery will take long time.
Tue Mar 14 16:26:29 2017 - Searching new master from slaves..
Tue Mar 14 16:26:29 2017 - Candidate masters from the configuration file:
Tue Mar 14 16:26:29 2017 -    mysql2(192.168.110.132:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 16:26:29 2017 -    GTID ON
Tue Mar 14 16:26:29 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 16:26:29 2017 -    Primary candidate for the new Master (candidate_master is set)
Tue Mar 14 16:26:29 2017 - Non-candidate masters:
Tue Mar 14 16:26:29 2017 -    mysql2(192.168.110.132:3306)Version=5.6.34-log (oldest major version between slaves) log-bin:enabled
Tue Mar 14 16:26:29 2017 -    GTID ON
Tue Mar 14 16:26:29 2017 -    Replicating from 192.168.110.131(192.168.110.131:3306)
Tue Mar 14 16:26:29 2017 -    Primary candidate for the new Master (candidate_master is set)
Tue Mar 14 16:26:29 2017 - Searching from candidate_master slaves which have received the latest relay log events..
Tue Mar 14 16:26:29 2017 -    Not found.
Tue Mar 14 16:26:29 2017 - Searching from all candidate_master slaves..
Tue Mar 14 16:26:29 2017 -    Not found.
Tue Mar 14 16:26:29 2017 - Searching from all slaves which have received the latest relay log events..
Tue Mar 14 16:26:29 2017 - New master is ansible(192.168.110.130:3306)
Tue Mar 14 16:26:29 2017 - Starting master failover..
Tue Mar 14 16:26:29 2017 -
From:
mysql1(192.168.110.131:3306) (current master)
+--mysql2(192.168.110.132:3306)
+--ansible(192.168.110.130:3306)

To:
ansible(192.168.110.130:3306) (new master)
+--mysql2(192.168.110.132:3306)
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - * Phase 3.3: New Master Recovery Phase..
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - Waiting all logs to be applied..
Tue Mar 14 16:26:29 2017 -    done.
Tue Mar 14 16:26:29 2017 - -- Saving binlog from host mysql2 started, pid: 39734
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - Log messages from mysql2 ...
Tue Mar 14 16:26:29 2017 -
Tue Mar 14 16:26:29 2017 - Fetching binary logs from binlog server mysql2..
Tue Mar 14 16:26:29 2017 - Executing binlog save command: save_binary_logs --command=save --start_file=mysql-bin.000001--start_pos=133309374 --output_file=/tmp/saved_binlog_binlog1_20170314162628.binlog --handle_raw_binlog=0 --skip_filter=1 --disable_log_bin=0 --manager_version=0.56 --oldest_version=5.6.34-log--binlog_dir=/tmp/binlog
Creating /tmp if not exists..    ok.
Concat binary/relay logs from mysql-bin.000001 pos 133309374 to mysql-bin.000015 EOF into /tmp/saved_binlog_binlog1_20170314162628.binlog ..
Target file /tmp/binlog/mysql-bin.000001 size=92195648, this is smaller than exec pos 133309374. Skipping.
Event not exists.
Tue Mar 14 16:26:29 2017 - Additional events were not found from the binlog server. No need to save.
Tue Mar 14 16:26:29 2017 - End of log messages from mysql2.
Tue Mar 14 16:26:29 2017 - No binlog events found from mysql2. Skipping
Tue Mar 14 16:26:29 2017 - Getting new master's binlog name and position..
Tue Mar 14 16:26:29 2017 - mysql-bin.000001:133306806
Tue Mar 14 16:26:29 2017 - All other slaves should start replication from here. Statement should be: CHANGE MASTER TO MASTER_HOST='ansible or 192.168.110.130', MASTER_PORT=3306, MASTER_AUTO_POSITION=1, MASTER_USER='repl', MASTER_PASSWORD='xxx';
Tue Mar 14 16:26:29 2017 - Master Recovery succeeded. File:Pos:Exec_Gtid_Set: mysql-bin.000001, 133306806, c897448e-07c8-11e7-8af3-000c29918f29:1-244
Tue Mar 14 16:26:29 2017 - Executing master IP activate script:
Tue Mar 14 16:26:29 2017 -    /usr/bin/master_ip_failover --command=start --ssh_user=root --orig_master_host=mysql1 --orig_master_ip=192.168.110.131 --orig_master_port=3306 --new_master_host=ansible --new_master_ip=192.168.110.130 --new_master_port=3306 --new_master_user='mha' --new_master_password='mha'
Unknown option: new_master_user
Unknown option: new_master_password


IN SCRIPT TEST====/sbin/ifconfig eth0:2 down==/sbin/ifconfig eth0:2 192.168.110.100;/sbin/arping -I eth0 -c 3 -s 192.168.110.100 192.168.110.2 >/dev/null 2>&1===

Enabling the VIP - 192.168.110.100 on the new master - ansible
Tue Mar 14 16:26:32 2017 - OK.
Tue Mar 14 16:26:33 2017 - ** Finished master recovery successfully.
Tue Mar 14 16:26:33 2017 - * Phase 3: Master Recovery Phase completed.
Tue Mar 14 16:26:33 2017 -
Tue Mar 14 16:26:33 2017 - * Phase 4: Slaves Recovery Phase..
Tue Mar 14 16:26:33 2017 -
Tue Mar 14 16:26:33 2017 -
Tue Mar 14 16:26:33 2017 - * Phase 4.1: Starting Slaves in parallel..
Tue Mar 14 16:26:33 2017 -
Tue Mar 14 16:26:33 2017 - -- Slave recovery on host mysql2(192.168.110.132:3306) started, pid: 39751. Check tmp log /var/log/masterha/app1/mysql2_3306_20170314162628.log if it takes time..
Tue Mar 14 16:27:14 2017 -
Tue Mar 14 16:27:14 2017 - Log messages from mysql2 ...
Tue Mar 14 16:27:14 2017 -
Tue Mar 14 16:26:33 2017 - Resetting slave mysql2(192.168.110.132:3306) and starting replication from the new master ansible(192.168.110.130:3306)..
Tue Mar 14 16:26:33 2017 - Executed CHANGE MASTER.
Tue Mar 14 16:26:33 2017 - Slave started.
Tue Mar 14 16:27:14 2017 - gtid_wait(c897448e-07c8-11e7-8af3-000c29918f29:1-244) completed on mysql2(192.168.110.132:3306). Executed 168 events.
Tue Mar 14 16:27:14 2017 - End of log messages from mysql2.
Tue Mar 14 16:27:14 2017 - -- Slave on host mysql2(192.168.110.132:3306) started.
Tue Mar 14 16:27:14 2017 - All new slave servers recovered successfully.
Tue Mar 14 16:27:14 2017 -
Tue Mar 14 16:27:14 2017 - * Phase 5: New master cleanup phase..
Tue Mar 14 16:27:14 2017 -
Tue Mar 14 16:27:14 2017 - Resetting slave info on the new master..
Tue Mar 14 16:27:15 2017 - ansible: Resetting slave info succeeded.
Tue Mar 14 16:27:15 2017 - Master failover to ansible(192.168.110.130:3306) completed successfully.
Tue Mar 14 16:27:15 2017 -

----- Failover Report -----

app1: MySQL Master failover mysql1(192.168.110.131:3306) to ansible(192.168.110.130:3306) succeeded

Master mysql1(192.168.110.131:3306) is down!

Check MHA Manager logs at ansible:/var/log/masterha/app1/manager.log for details.

Started automated(non-interactive) failover.
Invalidated master IP address on mysql1(192.168.110.131:3306)
Selected ansible(192.168.110.130:3306) as a new master.
ansible(192.168.110.130:3306): OK: Applying all logs succeeded.
ansible(192.168.110.130:3306): OK: Activated master IP address.
mysql2(192.168.110.132:3306): OK: Slave started, replicating from ansible(192.168.110.130:3306)
ansible(192.168.110.130:3306): Resetting slave info succeeded.
Master failover to ansible(192.168.110.130:3306) completed successfully.

查看完整版本: mysql5.6基于GTID模式之高可用架构搭建-MHA（mha0.56）