AI 助理
备案控制台
开发者社区 开发与运维 文章 正文

MFS+Keepalived双机高可用热备方案操作记录

2017-11-16 1297
版权
版权声明：
本文内容由阿里云实名注册用户自发贡献，版权归原作者所有，阿里云开发者社区不拥有其著作权，亦不承担相应法律责任。具体规则请查看《 阿里云开发者社区用户服务协议》和 《阿里云开发者社区知识产权保护指引》。如果您发现本社区中有涉嫌抄袭的内容，填写 侵权投诉表单进行举报，一经查实，本社区将立刻删除涉嫌侵权内容。
简介：

基于MFS的单点及手动备份的缺陷，考虑将其与Keepalived相结合以提高可用性。在Centos下MooseFS（MFS）分布式存储共享环境部署记录这篇文档部署环境的基础上，只需要做如下改动：

1
2
3
4
5
1）将master-server作为Keepalived_MASTER（启动mfsmaster、mfscgiserv）
2）将matelogger作为Keepalived_BACKUP（启动mfsmaster、mfscgiserv）
3）将ChunkServer服务器里配置的MASTER_HOST参数值改为VIP地址
4）clinet挂载的master的ip地址改为VIP地址
按照这样调整后，需要将Keepalived_MASTER和Keepalived_BACKUP里面的hosts绑定信息也修改下。

方案实施原理及思路

1
2
3
4
5
1）mfsmaster的故障恢复在1.6.5版本后可以由mfsmetalogger产生的日志文件changelog_ml.*.mfs和metadata.mfs.back文件通过命令mfsmetarestore恢复
2）定时从mfsmaster获取metadata.mfs.back 文件用于master恢复
3）Keepalived MASTER检测到mfsmaster进程宕停时会执行监控脚本，即自动启动mfsmaster进程，如果启动失败，则会强制 kill 掉keepalived和mfscgiserv进程，由此转移VIP到BACKUP上面。
4）若是Keepalived MASTER故障恢复，则会将VIP资源从BACKUP一方强制抢夺回来，继而由它提供服务
5）整个切换在2~5秒内完成 根据检测时间间隔。

架构拓扑图如下：

1）Keepalived_MASTER（mfs master）机器上的操作

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
mfs的master日志服务器的安装配置已经在另一篇文档中详细记录了，这里就不做赘述了。下面直接记录下keepalived的安装配置：
-----------------------------------------------------------------------------------------------------------------------
安装Keepalived
[root@Keepalived_MASTER ~] # yum install -y openssl-devel popt-devel
[root@Keepalived_MASTER ~] # cd /usr/local/src/
[root@Keepalived_MASTER src] # wget http://www.keepalived.org/software/keepalived-1.3.5.tar.gz
[root@Keepalived_MASTER src] # tar -zvxf keepalived-1.3.5.tar.gz
[root@Keepalived_MASTER src] # cd keepalived-1.3.5
[root@Keepalived_MASTER keepalived-1.3.5] # ./configure --prefix=/usr/local/keepalived
[root@Keepalived_MASTER keepalived-1.3.5] # make && make install
      
[root@Keepalived_MASTER keepalived-1.3.5] # cp /usr/local/src/keepalived-1.3.5/keepalived/etc/init.d/keepalived /etc/rc.d/init.d/
[root@Keepalived_MASTER keepalived-1.3.5] # cp /usr/local/keepalived/etc/sysconfig/keepalived /etc/sysconfig/
[root@Keepalived_MASTER keepalived-1.3.5] # mkdir /etc/keepalived/
[root@Keepalived_MASTER keepalived-1.3.5] # cp /usr/local/keepalived/etc/keepalived/keepalived.conf /etc/keepalived/
[root@Keepalived_MASTER keepalived-1.3.5] # cp /usr/local/keepalived/sbin/keepalived /usr/sbin/
[root@Keepalived_MASTER keepalived-1.3.5] # echo "/etc/init.d/keepalived start" >> /etc/rc.local
       
[root@Keepalived_MASTER keepalived-1.3.5] # chmod +x /etc/rc.d/init.d/keepalived      #添加执行权限
[root@Keepalived_MASTER keepalived-1.3.5] # chkconfig keepalived on                   #设置开机启动
[root@Keepalived_MASTER keepalived-1.3.5] # service keepalived start                   #启动
[root@Keepalived_MASTER keepalived-1.3.5] # service keepalived stop                    #关闭
[root@Keepalived_MASTER keepalived-1.3.5] # service keepalived restart                 #重启
     
配置Keepalived
[root@Keepalived_MASTER ~] # cp /etc/keepalived/keepalived.conf /etc/keepalived/keepalived.conf-bak
[root@Keepalived_MASTER ~] # vim /etc/keepalived/keepalived.conf
! Configuration File  for  keepalived
global_defs {
   notification_email {
     root@localhost
     }
   
notification_email_from keepalived@localhost
smtp_server 127.0.0.1
smtp_connect_timeout 30
router_id MFS_HA_MASTER
}
   
vrrp_script chk_mfs {                          
   script  "/usr/local/mfs/keepalived_check_mfsmaster.sh"
   interval 2
   weight 2
}
   
vrrp_instance VI_1 {
   state MASTER
   interface eth0
   virtual_router_id 51
   priority 100
   advert_int 1
   authentication {
     auth_type PASS
     auth_pass 1111
}
   track_script {
     chk_mfs
}
virtual_ipaddress {
     182.48.115.239
}
notify_master  "/etc/keepalived/clean_arp.sh 182.48.115.239"
}
     
     
接着编写监控脚本
[root@Keepalived_MASTER ~] # vim /usr/local/mfs/keepalived_check_mfsmaster.sh
#!/bin/bash
A=` ps  -C mfsmaster --no-header |  wc  -l`
if  [ $A - eq  0 ]; then
/etc/init .d /mfsmaster  start
sleep  3
    if  [ ` ps  -C mfsmaster --no-header |  wc  -l ` - eq  0 ]; then
       /usr/bin/killall  -9 mfscgiserv
       /usr/bin/killall  -9 keepalived
    fi
fi
 
[root@Keepalived_MASTER ~] # chmod 755 /usr/local/mfs/keepalived_check_mfsmaster.sh
 
设置更新虚拟服务器（VIP）地址的arp记录到网关脚本
[root@Keepalived_MASTER ~] # vim /etc/keepalived/clean_arp.sh
#!/bin/sh
VIP=$1
GATEWAY=182.48.115.254                                         // 这个是网关地址                      
/sbin/arping  -I eth0 -c 5 -s $VIP $GATEWAY &> /dev/null
 
[root@Keepalived_MASTER ~] # chmod 755 /etc/keepalived/clean_arp.sh
     
启动keepalived（确保Keepalived_MASTER机器的mfs master服务和Keepalived服务都要启动）
[root@Keepalived_MASTER ~] # /etc/init.d/keepalived start
Starting keepalived:                                       [  OK  ]
[root@Keepalived_MASTER ~] # ps -ef|grep keepalived
root     28718     1  0 13:09 ?        00:00:00 keepalived -D
root     28720 28718  0 13:09 ?        00:00:00 keepalived -D
root     28721 28718  0 13:09 ?        00:00:00 keepalived -D
root     28763 27466  0 13:09 pts /0     00:00:00  grep  keepalived
     
查看vip
[root@Keepalived_MASTER ~] # ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 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  52:54:00:09:21:60 brd ff:ff:ff:ff:ff:ff
     inet 182.48.115.233 /27  brd 182.48.115.255 scope global eth0
     inet 182.48.115.239 /32  scope global eth0
     inet6 fe80::5054:ff:fe09:2160 /64  scope link
        valid_lft forever preferred_lft forever

2）Keepalived_BACKUP（mfs master）机器上的操作

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
在另一篇文档里，该机器是作为metalogger元数据日志服务器的，那么在这个高可用环境下，该台机器改为Keepalived_BACKUP使用。
即去掉metalogger的部署，直接部署mfs master（部署过程参考另一篇文档）。下面直接说下Keepalived_BACKUP下的Keepalived配置：
  
安装Keepalived
[root@Keepalived_BACKUP ~] # yum install -y openssl-devel popt-devel
[root@Keepalived_BACKUP ~] # cd /usr/local/src/
[root@Keepalived_BACKUP src] # wget http://www.keepalived.org/software/keepalived-1.3.5.tar.gz
[root@Keepalived_BACKUP src] # tar -zvxf keepalived-1.3.5.tar.gz
[root@Keepalived_BACKUP src] # cd keepalived-1.3.5
[root@Keepalived_BACKUP keepalived-1.3.5] # ./configure --prefix=/usr/local/keepalived
[root@Keepalived_BACKUP keepalived-1.3.5] # make && make install
       
[root@Keepalived_BACKUP keepalived-1.3.5] # cp /usr/local/src/keepalived-1.3.5/keepalived/etc/init.d/keepalived /etc/rc.d/init.d/
[root@Keepalived_BACKUP keepalived-1.3.5] # cp /usr/local/keepalived/etc/sysconfig/keepalived /etc/sysconfig/
[root@Keepalived_BACKUP keepalived-1.3.5] # mkdir /etc/keepalived/
[root@Keepalived_BACKUP keepalived-1.3.5] # cp /usr/local/keepalived/etc/keepalived/keepalived.conf /etc/keepalived/
[root@Keepalived_BACKUP keepalived-1.3.5] # cp /usr/local/keepalived/sbin/keepalived /usr/sbin/
[root@Keepalived_BACKUP keepalived-1.3.5] # echo "/etc/init.d/keepalived start" >> /etc/rc.local
        
[root@Keepalived_BACKUP keepalived-1.3.5] # chmod +x /etc/rc.d/init.d/keepalived      #添加执行权限
[root@Keepalived_BACKUP keepalived-1.3.5] # chkconfig keepalived on                   #设置开机启动
[root@Keepalived_BACKUP keepalived-1.3.5] # service keepalived start                   #启动
[root@Keepalived_BACKUP keepalived-1.3.5] # service keepalived stop                    #关闭
[root@Keepalived_BACKUP keepalived-1.3.5] # service keepalived restart                 #重启
      
配置Keepalived
[root@Keepalived_BACKUP ~] # cp /etc/keepalived/keepalived.conf /etc/keepalived/keepalived.conf-bak
[root@Keepalived_BACKUP ~] # vim /etc/keepalived/keepalived.conf
! Configuration File  for  keepalived
global_defs {
   notification_email {
     root@localhost
     }
   
notification_email_from keepalived@localhost
smtp_server 127.0.0.1
smtp_connect_timeout 30
router_id MFS_HA_BACKUP
}
   
vrrp_script chk_mfs {                          
   script  "/usr/local/mfs/keepalived_check_mfsmaster.sh"
   interval 2
   weight 2
}
   
vrrp_instance VI_1 {
   state BACKUP
   interface eth0
   virtual_router_id 51
   priority 99
   advert_int 1
   authentication {
     auth_type PASS
     auth_pass 1111
}
   track_script {
     chk_mfs
}
virtual_ipaddress {
     182.48.115.239
}
notify_master  "/etc/keepalived/clean_arp.sh 182.48.115.239"
}
  
  
接着编写监控脚本
[root@Keepalived_BACKUP ~] # vim /usr/local/mfs/keepalived_notify.sh
#!/bin/bash
A=` ps  -C mfsmaster --no-header |  wc  -l`
if  [ $A - eq  0 ]; then
/etc/init .d /mfsmaster  start
sleep  3
    if  [ ` ps  -C mfsmaster --no-header |  wc  -l ` - eq  0 ]; then
       /usr/bin/killall  -9 mfscgiserv
       /usr/bin/killall  -9 keepalived
    fi
fi
  
[root@Keepalived_BACKUP ~] # chmod 755 /usr/local/mfs/keepalived_notify.sh
 
设置更新虚拟服务器（VIP）地址的arp记录到网关脚本（Haproxy_Keepalived_Master 和 Haproxy_Keepalived_Backup两台机器都要操作）
[root@Keepalived_BACKUP ~] # vim /etc/keepalived/clean_arp.sh
#!/bin/sh
VIP=$1
GATEWAY=182.48.115.254                                     
/sbin/arping  -I eth0 -c 5 -s $VIP $GATEWAY &> /dev/null
  
启动keepalived
[root@Keepalived_BACKUP ~] # /etc/init.d/keepalived start
Starting keepalived:
[root@Keepalived_BACKUP ~] # ps -ef|grep keepalived
root     17565     1  0 11:06 ?        00:00:00 keepalived -D
root     17567 17565  0 11:06 ?        00:00:00 keepalived -D
root     17568 17565  0 11:06 ?        00:00:00 keepalived -D
root     17778 17718  0 13:47 pts /1     00:00:00  grep  keepalived
  
要保证Keepalived_BACKUP机器的mfs master服务和keepalived服务都要启动！

3）chunkServer的配置

1
2
只需要将mfschunkserver.cfg文件中的MASTER_HOST参数配置成182.48.115.239，即VIP地址。
其他的配置都不需要修改。然后重启mfschunkserver服务

4）clinet客户端的配置

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
只需要见挂载命令中的元数据ip改为182.48.115.249即可！
 
[root@clinet-server ~] # mkdir /mnt/mfs               
[root@clinet-server ~] # mkdir /mnt/mfsmeta
 
[root@clinet-server ~] # /usr/local/mfs/bin/mfsmount /mnt/mfs -H 182.48.115.239
mfsmaster accepted connection with parameters:  read -write,restricted_ip,admin ; root mapped to root:root
 
[root@clinet-server ~] # /usr/local/mfs/bin/mfsmount -m /mnt/mfsmeta -H 182.48.115.239
mfsmaster accepted connection with parameters:  read -write,restricted_ip
 
[root@clinet-server ~] # df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/VolGroup-lv_root
                       8.3G  3.8G  4.1G  49% /
tmpfs                 499M  228K  498M   1%  /dev/shm
/dev/vda1              477M   35M  418M   8%  /boot
/dev/sr0               3.7G  3.7G     0 100%  /media/CentOS_6 .8_Final
182.48.115.239:9421   107G   42G   66G  39%  /mnt/mfs
 
[root@clinet-server ~] # mount
/dev/mapper/VolGroup-lv_root  on /  type  ext4 (rw)
proc on  /proc  type  proc (rw)
sysfs on  /sys  type  sysfs (rw)
devpts on  /dev/pts  type  devpts (rw,gid=5,mode=620)
tmpfs on  /dev/shm  type  tmpfs (rw,rootcontext= "system_u:object_r:tmpfs_t:s0" )
/dev/vda1  on  /boot  type  ext4 (rw)
none on  /proc/sys/fs/binfmt_misc  type  binfmt_misc (rw)
gvfs-fuse-daemon on  /root/ .gvfs  type  fuse.gvfs-fuse-daemon (rw,nosuid,nodev)
/dev/sr0  on  /media/CentOS_6 .8_Final  type  iso9660 (ro,nosuid,nodev,uhelper=udisks,uid=0,gid=0,iocharset=utf8,mode=0400,dmode=0500)
182.48.115.239:9421 on  /mnt/mfs  type  fuse.mfs (rw,nosuid,nodev,allow_other)
182.48.115.239:9421 on  /mnt/mfsmeta  type  fuse.mfsmeta (rw,nosuid,nodev,allow_other)
 
验证下客户端挂载MFS文件系统后的数据读写是否正常
[root@clinet-server ~] # cd /mnt/mfs
[root@clinet-server mfs] # echo "12312313" > test.txt
[root@clinet-server mfs] # cat test.txt
12312313
 
[root@clinet-server mfs] # rm -f test.txt
[root@clinet-server mfs] # cd ../mfsmeta/trash/
[root@clinet-server trash] # find . -name "*test*"
. /003/00000003 | test .txt
[root@clinet-server trash] # cd ./003/
[root@clinet-server 003] # ls
00000003| test .txt  undel
[root@clinet-server 003] # mv 00000003\|test.txt undel/
[root@clinet-server 003] # ls /mnt/mfs
test .txt
[root@clinet-server 003] # cat /mnt/mfs/test.txt
12312313
 
以上说明挂载后的MFS数据共享正常

5）Keepalived_MASTER和Keepalived_BACKUP的iptales防火墙设置

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Keepalived_MASTER和Keepalived_BACKUP的防火墙iptables在实验中是关闭的。
如果开启了iptables防火墙功能，则需要在两台机器的iptables里配置如下：
 
可以使用 "ss -l" 命令查看本机监听的端口
[root@Keepalived_MASTER ~] # ss -l
State       Recv-Q Send-Q                                     Local Address:Port                                         Peer Address:Port  
LISTEN      0      100                                                    *:9419                                                    *:*      
LISTEN      0      100                                                    *:9420                                                    *:*      
LISTEN      0      100                                                    *:9421                                                    *:*      
LISTEN      0      50                                                     *:9425                                                    *:*      
LISTEN      0      128                                                   ::: ssh                                                     :::*      
LISTEN      0      128                                                    *: ssh                                                      *:*      
LISTEN      0      100                                                  ::1:smtp                                                   :::*      
LISTEN      0      100                                            127.0.0.1:smtp                                                    *:*   
 
[root@Keepalived_MASTER ~] # vim /etc/sysconfig/iptables
........
-A INPUT -s 182.148.15.0 /24  -d 224.0.0.18 -j ACCEPT        #允许组播地址通信。注意设置这两行，就会在Keepalived_MASTER故障恢复后，将VIP资源从Keepalived_BACK那里再转移回来
-A INPUT -s 182.148.15.0 /24  -p vrrp -j ACCEPT              #允许VRRP（虚拟路由器冗余协）通信
-A INPUT -m state --state NEW -m tcp -p tcp --dport 9419 -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 9420 -j ACCEPT 
-A INPUT -m state --state NEW -m tcp -p tcp --dport 9421 -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 9425 -j ACCEPT
 
[root@Keepalived_MASTER ~] # /etc/init.d/iptables start

6）故障切换后的数据同步脚本

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
上面的配置可以实现Keepalived_MASTER机器出现故障（keepalived服务关闭），VIP资源转移到Keepalived_BACKUP上；
当Keepalived_MASTER机器故障恢复（即keepalived服务开启），那么它就会将VIP资源再次抢夺回来！
 
但是只是实现了VIP资源的转移，但是MFS文件系统的数据该如何进行同步呢？
下面在两台机器上分别写了数据同步脚本（Keepalived_MASTER和Keepalived_BACKUP要提前做好双方的 ssh 无密码登陆的信任关系）
 
Keepalived_MASTER机器上
[root@Keepalived_MASTER ~] # vim /usr/local/mfs/MFS_DATA_Sync.sh
#!/bin/bash
A=`ip addr| grep  182.48.115.239| awk  -F " "  '{print $2}' | cut  -d "/"  -f1`
if  [ $A == 182.48.115.239 ]; then
    /etc/init .d /mfsmaster  stop
    /bin/rm  -f  /usr/local/mfs/var/mfs/ *
    /usr/bin/rsync  -e  "ssh -p22"  -avpgolr 182.48.115.235: /usr/local/mfs/var/mfs/ *  /usr/local/mfs/var/mfs/
    /usr/local/mfs/sbin/mfsmetarestore  -m
    /etc/init .d /mfsmaster  -a
    sleep  3
    echo  "this server has become the master of MFS"
    if  [ $A != 182.48.115.239 ]; then
    echo  "this server is still MFS's slave"
    fi
fi
 
Keepalived_BACKUP机器上
[root@Keepalived_BACKUP ~] # vim /usr/local/mfs/MFS_DATA_Sync.sh
#!/bin/bash
A=`ip addr| grep  182.48.115.239| awk  -F " "  '{print $2}' | cut  -d "/"  -f1`
if  [ $A == 182.48.115.239 ]; then
    /etc/init .d /mfsmaster  stop
    /bin/rm  -f  /usr/local/mfs/var/mfs/ *
    /usr/bin/rsync  -e  "ssh -p22"  -avpgolr 182.48.115.233: /usr/local/mfs/var/mfs/ *  /usr/local/mfs/var/mfs/
    /usr/local/mfs/sbin/mfsmetarestore  -m
    /etc/init .d /mfsmaster  -a
    sleep  3
    echo  "this server has become the master of MFS"
    if  [ $A != 182.48.115.239 ]; then
    echo  "this server is still MFS's slave"
    fi
fi
 
即当VIP资源转移到自己这一方时，执行这个同步脚本，就会将对方的数据同步过来了。

7）故障切换测试

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
1）关闭Keepalived_MASTER的mfsmaster服务
由于keepalived.conf文件中的监控脚本定义，当发现mfsmaster进程不存在时，就会主动启动mfsmaster。只要当mfsmaster启动失败，才会强制
killall掉keepalived和mfscgiserv进程
 
[root@Keepalived_MASTER ~] # /etc/init.d/mfsmaster stop
sending SIGTERM to lock owner (pid:29266)
waiting  for  termination terminated
 
发现mfsmaster关闭后，会自动重启
[root@Keepalived_MASTER ~] # ps -ef|grep mfs
root     26579     1  0 16:00 ?        00:00:00  /usr/bin/python  /usr/local/mfs/sbin/mfscgiserv  start
root     30389 30388  0 17:18 ?        00:00:00  /bin/bash  /usr/local/mfs/keepalived_check_mfsmaster .sh
mfs      30395     1 71 17:18 ?        00:00:00  /etc/init .d /mfsmaster  start
 
默认情况下，VIP资源是在Keepalived_MASTER上的
[root@Keepalived_MASTER ~] # ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 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  52:54:00:09:21:60 brd ff:ff:ff:ff:ff:ff
     inet 182.48.115.233 /27  brd 182.48.115.255 scope global eth0
     inet 182.48.115.239 /32  scope global eth0
     inet6 fe80::5054:ff:fe09:2160 /64  scope link
        valid_lft forever preferred_lft forever
 
在client端挂载后（通过VIP地址挂载），查看数据
[root@clinet-server ~] # cd /mnt/mfs
[root@clinet-server mfs] # ll
total 4
-rw-r--r-- 1 root root 9 May 24 17:11 grace
-rw-r--r-- 1 root root 9 May 24 17:11 grace1
-rw-r--r-- 1 root root 9 May 24 17:11 grace2
-rw-r--r-- 1 root root 9 May 24 17:11 grace3
-rw-r--r-- 1 root root 9 May 24 17:10 kevin
-rw-r--r-- 1 root root 9 May 24 17:10 kevin1
-rw-r--r-- 1 root root 9 May 24 17:10 kevin2
-rw-r--r-- 1 root root 9 May 24 17:10 kevin3
 
当keepalived关闭（这个时候mfsmaster关闭后就不会自动重启了，因为keepalived关闭了，监控脚本就不会执行了）
[root@Keepalived_MASTER ~] # /etc/init.d/keepalived stop
Stopping keepalived:                                       [  OK  ]
[root@Keepalived_MASTER ~] # ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 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  52:54:00:09:21:60 brd ff:ff:ff:ff:ff:ff
     inet 182.48.115.233 /27  brd 182.48.115.255 scope global eth0
     inet6 fe80::5054:ff:fe09:2160 /64  scope link
        valid_lft forever preferred_lft forever
 
发现，Keepalived_MASTER的keepalived关闭后，VIP资源就不在它上面了。
 
查看系统日志，发现VIP已经转移
[root@Keepalived_MASTER ~] # tail -1000 /var/log/messages
.......
May 24 17:11:19 centos6-node1 Keepalived_vrrp[29184]: VRRP_Instance(VI_1) Sending /queueing  gratuitous ARPs on eth0  for  182.48.115.239
May 24 17:11:19 centos6-node1 Keepalived_vrrp[29184]: Sending gratuitous ARP on eth0  for  182.48.115.239
May 24 17:11:19 centos6-node1 Keepalived_vrrp[29184]: Sending gratuitous ARP on eth0  for  182.48.115.239
May 24 17:11:19 centos6-node1 Keepalived_vrrp[29184]: Sending gratuitous ARP on eth0  for  182.48.115.239
May 24 17:11:19 centos6-node1 Keepalived_vrrp[29184]: Sending gratuitous ARP on eth0  for  182.48.115.239
 
 
然后到Keepalived_BACKUP上面发现，VIP已经过来了
[root@Keepalived_BACKUP ~] # ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 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  52:54:00:82:69:69 brd ff:ff:ff:ff:ff:ff
     inet 182.48.115.235 /27  brd 182.48.115.255 scope global eth0
     inet 182.48.115.239 /32  scope global eth0
     inet6 fe80::5054:ff:fe82:6969 /64  scope link
        valid_lft forever preferred_lft forever
 
查看日志，也能看到VIP转移过来了
[root@Keepalived_BACKUP ~] # tail -1000 /var/log/messages
.......
May 24 17:27:57 centos6-node2 Keepalived_vrrp[5254]: VRRP_Instance(VI_1) Received advert with higher priority 100, ours 99
May 24 17:27:57 centos6-node2 Keepalived_vrrp[5254]: VRRP_Instance(VI_1) Entering BACKUP STATE
May 24 17:27:57 centos6-node2 Keepalived_vrrp[5254]: VRRP_Instance(VI_1) removing protocol VIPs.
 
再次在clinet客户端挂载后，查看数据
[root@clinet-server mfs] # ll
[root@clinet-server mfs] #
 
发现没有数据，这就需要执行上面的那个同步脚本
[root@Keepalived_BACKUP ~] # sh -x /usr/local/mfs/MFS_DATA_Sync.sh
 
再次在clinet客户端挂载后，查看数据
[root@clinet-server mfs] # ll
total 4
-rw-r--r--. 1 root root 9 May 24 17:11 grace
-rw-r--r--. 1 root root 9 May 24 17:11 grace1
-rw-r--r--. 1 root root 9 May 24 17:11 grace2
-rw-r--r--. 1 root root 9 May 24 17:11 grace3
-rw-r--r--. 1 root root 9 May 24 17:10 kevin
-rw-r--r--. 1 root root 9 May 24 17:10 kevin1
-rw-r--r--. 1 root root 9 May 24 17:10 kevin2
-rw-r--r--. 1 root root 9 May 24 17:10 kevin3
 
发现数据已经同步过来，然后再更新数据
[root@clinet-server mfs] # rm -f ./*
[root@clinet-server mfs] # echo "123123" > wangshibo
[root@clinet-server mfs] # echo "123123" > wangshibo1
[root@clinet-server mfs] # echo "123123" > wangshibo2
[root@clinet-server mfs] # echo "123123" > wangshibo3
[root@clinet-server mfs] # echo "123123" > wangshibo4
[root@clinet-server mfs] # ll
total 3
-rw-r--r--. 1 root root 7 May 24 17:26 wangshibo
-rw-r--r--. 1 root root 7 May 24 17:26 wangshibo1
-rw-r--r--. 1 root root 7 May 24 17:26 wangshibo2
-rw-r--r--. 1 root root 7 May 24 17:26 wangshibo3
-rw-r--r--. 1 root root 7 May 24 17:26 wangshibo4
 
 
2）恢复Keepalived_MASTER的keepalived进程
[root@Keepalived_MASTER ~] # /etc/init.d/keepalived start
Starting keepalived:                                       [  OK  ]
[root@Keepalived_MASTER ~] # ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 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  52:54:00:09:21:60 brd ff:ff:ff:ff:ff:ff
     inet 182.48.115.233 /27  brd 182.48.115.255 scope global eth0
     inet 182.48.115.239 /32  scope global eth0
     inet6 fe80::5054:ff:fe09:2160 /64  scope link
        valid_lft forever preferred_lft forever
 
发现Keepalived_MASTER的keepalived进程启动后，VIP资源又抢夺回来。查看 /var/log/messages 日志能看出VIP资源转移回来
 
再次查看Keepalived_BACKUP，发现VIP不在了
[root@Keepalived_BACKUP ~] # ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 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  52:54:00:82:69:69 brd ff:ff:ff:ff:ff:ff
     inet 182.48.115.235 /27  brd 182.48.115.255 scope global eth0
     inet6 fe80::5054:ff:fe82:6969 /64  scope link
        valid_lft forever preferred_lft forever
 
在client端挂载后（通过VIP地址挂载），查看数据
[root@clinet-server mfs] # ll
total 4
-rw-r--r--. 1 root root 9 May 24 17:11 grace
-rw-r--r--. 1 root root 9 May 24 17:11 grace1
-rw-r--r--. 1 root root 9 May 24 17:11 grace2
-rw-r--r--. 1 root root 9 May 24 17:11 grace3
-rw-r--r--. 1 root root 9 May 24 17:10 kevin
-rw-r--r--. 1 root root 9 May 24 17:10 kevin1
-rw-r--r--. 1 root root 9 May 24 17:10 kevin2
-rw-r--r--. 1 root root 9 May 24 17:10 kevin3
 
发现数据还是旧的，需要在Keepalived_MASTER上执行同步
[root@Keepalived_MASTER ~] # sh -x /usr/local/mfs/MFS_DATA_Sync.sh
 
再次在clinet客户端挂载后，查看数据,发现已经同步
[root@xqsj_web3 mfs] # ll
total 3
-rw-r--r-- 1 root root 7 May 24 17:26 wangshibo
-rw-r--r-- 1 root root 7 May 24 17:26 wangshibo1
-rw-r--r-- 1 root root 7 May 24 17:26 wangshibo2
-rw-r--r-- 1 root root 7 May 24 17:26 wangshibo3
-rw-r--r-- 1 root root 7 May 24 17:26 wangshibo4
***************当你发现自己的才华撑不起野心时，就请安静下来学习吧***************

本文转自散尽浮华博客园博客，原文链接：http://www.cnblogs.com/kevingrace/p/5570604.html ，如需转载请自行联系原作者
文章标签：
网络协议
开发工具
Shell
网络安全
监控
吞吞吐吐的
目录
相关文章
心跳包
|
Linux
LINUX进阶（其他篇）之make编译报错原因
LINUX进阶（其他篇）之make编译报错原因
心跳包
552 0
嘟嘟嘟嘟嘟嘟
|
运维 关系型数据库 MySQL
PolarDB产品使用问题之PolarDB MySQL版和PolarDB-X的区别是什么
PolarDB产品使用合集涵盖了从创建与管理、数据管理、性能优化与诊断、安全与合规到生态与集成、运维与支持等全方位的功能和服务，旨在帮助企业轻松构建高可用、高性能且易于管理的数据库环境，满足不同业务场景的需求。用户可以通过阿里云控制台、API、SDK等方式便捷地使用这些功能，实现数据库的高效运维与持续优化。
嘟嘟嘟嘟嘟嘟
465 1
wljslmz
|
SQL 数据管理 数据库
DROP 和 TRUNCATE 命令的详细区别
【8月更文挑战第31天】
wljslmz
1521 0
艾利克斯冰
|
Java
Byte 高位/低位简介绍（大端格式/小端格式）
Byte 高位/低位简介绍（大端格式/小端格式）
艾利克斯冰
1896 1
jjczqyjxgdegi
|
机器学习/深度学习 关系型数据库 Shell
老程序员分享：Linux之pushd、popd和dirs的使用讲解，比cd
老程序员分享：Linux之pushd、popd和dirs的使用讲解，比cd
jjczqyjxgdegi
305 1
DataBlog
|
供应链
2023湖北省中小企业数字化转型试点城市名单
黄石市、襄阳市、荆州市、十堰市、荆门市等5个城市为首批湖北省中小企业数字化转型试点城市。
DataBlog
491 1
上云小帮手
|
弹性计算 应用服务中间件 Apache
2核4g云服务器支持多少人同时在线？2核4G服务器并发数计算
阿里云服务器2核4g并发数是多少？支持多少人同时在线？
上云小帮手
1349 0
2核4g云服务器支持多少人同时在线？2核4G服务器并发数计算
云栖号资讯小哥
|
SQL 存储 分布式计算
菜鸟数据中台技术演进之路
通过中台的各技术域能力的建设，技术人员在极少的投入下，就可以支撑数倍的分析人员进行数字化运营工作。3 年时间里，菜鸟走过了从人力支撑到中台支撑的历程。
云栖号资讯小哥
4809 0
技术小甜
|
Linux
linux下最容易的增量备份,tar增量备份
技术小甜
2592 0
阿里云基础设施
|
存储 人工智能 数据中心
阿里云基础设施网络亮相SIGCOMM22 - 可预期网络取得重大突破
阿里云基础设施网络亮相SIGCOMM22 - 可预期网络取得重大突破
阿里云基础设施
1136 0
阿里云基础设施网络亮相SIGCOMM22 - 可预期网络取得重大突破

热门文章

最新文章

  • 1
    linux下ffmpeg安装
  • 2
    overleaf 插入图片，引用图片，图标标题Fig与文章引用Figure不一致解决
  • 3
    364 页 PyTorch 版《动手学深度学习》PDF 开源了（全中文，支持 Jupyter 运行）
  • 4
    阿里云上的几种数据备份方式
  • 5
    用深度神经网络修复H漫：看完这篇你就能眼中无码
  • 6
    MySQL · 引擎特性 · 像NOSQL那样使用MySQL
  • 7
    解决GitHub文件无法下载的问题
  • 8
    OpenSearch大模型实践之Havenask篇
  • 9
    SpringBoot整合Java Mail实现Outlook / Office365发送邮件
  • 10
    2019阿里云双11来袭,价格非常重要,但是懂得上云攻略更重要
  • 1
    AI Agent框架
    62
  • 2
    AI Agent：从工具到伙伴的智能进化
    72
  • 3
    计算机网络--TCP和UDP学习
    30
  • 4
    LBA-ECO ND-30 营养分析和气体通量，森林年代序列，巴西帕拉州
    30
  • 5
    2025年测试工程师的核心能力：掌握Dify工作流编排AI测试智能体
    60
  • 6
    有了解过 SpringBoot 的参数配置吗？
    53
  • 7
    分类模型校准：ROC-AUC不够？用ECE/pMAD评估概率质量
    24
  • 8
    把服务器当代码写：IaC 让部署变得又稳又香
    49
  • 9
    SDN 和 NFV：把网络“拆开重做”的底层革命，到底牛在哪？
    54
  • 10
    用数据“掘地三尺”，为零碳排放找出一条最靠谱的路
    32

    • 相关电子书

    更多
  • 低代码开发师（初级）实战教程
  • 冬季实战营第三期：MySQL数据库进阶实战
  • 阿里巴巴DevOps 最佳实践手册
    • 下一篇
    阿里云弹性公网IP线路类型【BGP(多线)_精品】是什么意思？