原地址：http://blog.csdn.net/moqiang02/article/details/37921051
概念简单认知：
Keepalived:它的诞生最初是为ipvs（一些服务，内核中的一些规则）提供高可用性的，最初最主要目的是能够自主调用ipvsadm来生成规则，并且能够自动实现将用户访问的地址转移到其他节点上进行实现的。
Keepalived:核心包含两个ckeckers和VRRP协议。
ckeckers：检查服务检查reserved的健康状况的，基于脚本也可检查服务本身的健康状况。这里是实现ipvs后端健康状况的检测的。
VRRP：是一种容错协议，它保证当主机的下一跳路由器出现故障时，由另一台路由器来代替出现故障的路由器进行工作，从而保持网络通信的连续性和可靠性。VRRP中每个节点之间都有优先级的一般为0-255（0,255有特殊用法）数字越大优先级越高。
相关术语解析：
虚拟路由器：由一个Master路由器和多个Backup路由器组成。主机将虚拟路由器当作默认网关。
VRID：虚拟路由器的标识。有相同VRID的一组路由器构成一个虚拟路由器。
Master路由器：虚拟路由器中承担报文转发任务的路由器。
Backup路由器：Master路由器出现故障时，能够代替Master路由器工作的路由器。
虚拟IP 地址：虚拟路由器的IP 地址。一个虚拟路由器可以拥有一个或多个IP地址。
IP地址拥有者：接口IP地址与虚拟IP地址相同的路由器被称为IP地址拥有者。
虚拟MAC地址：一个虚拟路由器拥有一个虚拟MAC地址。虚拟MAC地址的格式为00-00-5E-00-01-{VRID}。通常情况下，虚拟路由器回应ARP请求使用的是虚拟MAC地址，只有虚拟路由器做特殊配置的时候，才回应接口的真实MAC地址。
优先级：VRRP根据优先级来确定虚拟路由器中每台路由器的地位。
非抢占方式：如果Backup路由器工作在非抢占方式下，则只要Master路由器没有出现故障Backup路由器即使随后被配置了更高的优先级也不会成为Master路由器。
抢占方式：如果Backup路由器工作在抢占方式下，当它收到VRRP报文后，会将自己的优先级与通告报文中的优先级进行比较。如果自己的优先级比当前的Master路由器的优先级高，就会主动抢占成为Master路由器；否则，将保持Backup状态。
平台信息介绍：
Master：172.16.18.7
backup：172.16.18.9
系统版本：centosx86_64
keepalived版本：1.2.7
主配置文件：/etc/keepalived/keepalived.conf
服务脚本：/etc/rc.d/init,d/keepalived
应用实践：
将两个节点的时间同步
|
1
2
3
4
5
6
7
8
9
10
11
12
|
##############实现双机互信################node1#######ssh-keygen -t rsa -P ''ssh-copy-id -i .ssh/id_rsa.pub root@172.16.18.9#########node2#######ssh-keygen -t rsa -P ''ssh-copy-id -i .ssh/id_rsa.pub root@172.16.18.7##############查看时间###########[root@node1 ~]# date;ssh node2 'date'#####为实现同步可使用下面同步#####crontab -e*/5 * * * * /usr/sbin/ntpdate 172.16.0.1 &> /dev/null |
安装：
|
1
|
[root@node1 ~]# yum -y install keepalived |
查看编辑配置文件：
|
1
|
vim /etc/keepalived/keepalived.conf |
解析配置文件：
配置文件有三部分组成：
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
|
（1）：GLOBAL CONFIGURATION 全局配置段 有两个字段： Global definitions #全局定义 Static routes #静态路由（2）：VRRPD CONFIGURATION 配置VRRP子进程协议段又称定义虚拟路由的 有两个字段： VRRP synchronization group(s) #VRRP的同步组（一般不用） 什么是同步组？就是一台机器上有配置两个VIP，为了实现两个VIP要同步工作同时转移出去，所以必须要定义成同步组从而当成一个资源来转移。 VRRP instance(s) #VRRP的实例 任何一个虚拟路由定义好之后在任何一个节点上都应该定义一个keepalived运行实例，这两个节点上的实例要匹配。keepalived最令人头疼的是两个节点上的初始实例是不一样的，因为每一个节点都有初始状态而且它有默认的优先级，高的为Master低的为Backup所以导致了两个节点上的虚拟路由的实例配置是不一样的。（3）：LVS CONFIGURATION LVS配置段 有两个字段： Virtual server group(s) #虚拟服务器组 Virtual server(s) #虚拟服务器（ipvs规则） |
详细解析：
keepalived.conf
|
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
|
global_defs { #全局配置，这里额外的静态路由并未添加因为它是非必要的，除非我们在当前或特定的主机上生成特殊的静态路由等 notification_email { #收件人信息 acassen@firewall.loc failover@firewall.loc sysadmin@firewall.loc } notification_email_from Alexandre.Cassen@firewall.loc #发件人信息（可以随意伪装） smtp_server 192.168.200.1 #发邮件的服务器（一定不可为外部地址） smtp_connect_timeout 30 #连接超时时间 router_id LVS_DEVEL #路由器的标识（可以随便改动）}vrrp_instance VI_1 { #配置虚拟路由器的（VI_1是实例名称） state MASTER #初始状态，master|backup，当state指定的instance的初始化状态，在两台服务器都启动以后，马上发生竞选，优先级高的成为MASTER，所以这里的MASTER并不是表示此台服务器一直是MASTER interface eth0 #通告选举所用端口 virtual_router_id 51 #虚拟路由的ID号（一般不可大于255） priority 100 #优先级信息 advert_int 1 #初始化通告几个 authentication { #认证 auth_type PASS #认证机制 auth_pass 1111 #密码（尽量使用随机） } virtual_ipaddress { #虚拟地址（VIP地址） 192.168.200.16 192.168.200.17 192.168.200.18 }} |
编辑设置配置信息：
|
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
|
[root@node1 ~]# vim /etc/keepalived/keepalived.conf #主节点global_defs { notification_email { root@localhost } notification_email_from keadmin@localhost smtp_server 127.0.0.1 smtp_connect_timeout 30 router_id LVS_DEVEL}vrrp_instance VI_1 { state MASTER interface eth0 virtual_router_id 55 priority 100 advert_int 1 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 172.16.18.100 }}[root@node1 ~]# scp /etc/keepalived/keepalived.conf 172.16.18.9:/etc/keepalived/ #复制至备节点[root@node2 keepalived]# vim keepalived.conf #备节点global_defs { notification_email { root@localhost } notification_email_from keadmin@localhost smtp_server 127.0.0.1 smtp_connect_timeout 30 router_id LVS_DEVEL} vrrp_instance VI_1 { state BACKUP #状态 interface eth0 virtual_router_id 55 #一定要和主节点一致 priority 90 #优先级别一定低于主节点 advert_int 1 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 172.16.18.100 }} |
启动主节点：
|
1
|
[root@node1 ~]# service keepalived start |
启动备节点：
|
1
|
[root@node2 ~]# service keepalived start |
查看状态：
|
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
|
###############节点一############[root@node1 ~]# ip addr show1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:06:a6:49 brd ff:ff:ff:ff:ff:ff inet 172.16.18.7/16 brd 172.16.255.255 scope global eth0 inet 172.16.18.100/32 scope global eth0 #此时VIP在node1节点上 inet6 fe80::20c:29ff:fe06:a649/64 scope link valid_lft forever preferred_lft forever###############节点二############[root@node2 keepalived]# ip addr show1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:12:c8:b5 brd ff:ff:ff:ff:ff:ff inet 172.16.18.9/16 brd 172.16.255.255 scope global eth0 inet6 fe80::20c:29ff:fe12:c8b5/64 scope link valid_lft forever preferred_lft forever |
查看主节点启动日志信息：
|
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
|
##############主节点#############[root@node1 ~]# tail -20 /var/log/messagesSep 25 17:32:17 node1 Keepalived_healthcheckers[16628]: Netlink reflector reports IP 172.16.18.7 addedSep 25 17:32:17 node1 Keepalived_healthcheckers[16628]: Netlink reflector reports IP fe80::20c:29ff:fe06:a649 addedSep 25 17:32:17 node1 Keepalived_healthcheckers[16628]: Registering Kernel netlink reflectorSep 25 17:32:17 node1 Keepalived_healthcheckers[16628]: Registering Kernel netlink command channelSep 25 17:32:17 node1 Keepalived_healthcheckers[16628]: Opening file '/etc/keepalived/keepalived.conf'.Sep 25 17:32:17 node1 Keepalived_healthcheckers[16628]: Configuration is using : 6832 BytesSep 25 17:32:18 node1 kernel: IPVS: Registered protocols (TCP, UDP, SCTP, AH, ESP)Sep 25 17:32:18 node1 kernel: IPVS: Connection hash table configured (size=4096, memory=64Kbytes)Sep 25 17:32:18 node1 kernel: IPVS: ipvs loaded.Sep 25 17:32:18 node1 Keepalived_healthcheckers[16628]: Using LinkWatch kernel netlink reflector...Sep 25 17:32:18 node1 Keepalived_vrrp[16629]: Opening file '/etc/keepalived/keepalived.conf'.Sep 25 17:32:18 node1 Keepalived_vrrp[16629]: Configuration is using : 62657 BytesSep 25 17:32:18 node1 Keepalived_vrrp[16629]: Using LinkWatch kernel netlink reflector...Sep 25 17:32:18 node1 Keepalived_vrrp[16629]: VRRP sockpool: [ifindex(2), proto(112), fd(11,12)]Sep 25 17:32:19 node1 Keepalived_vrrp[16629]: VRRP_Instance(VI_1) Transition to MASTER STATE #事务开始转换为Master状态Sep 25 17:32:20 node1 Keepalived_vrrp[16629]: VRRP_Instance(VI_1) Entering MASTER STATE #进入master状态Sep 25 17:32:20 node1 Keepalived_vrrp[16629]: VRRP_Instance(VI_1) setting protocol VIPs.Sep 25 17:32:20 node1 Keepalived_vrrp[16629]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 172.16.18.100Sep 25 17:32:20 node1 Keepalived_healthcheckers[16628]: Netlink reflector reports IP 172.16.18.100 added #添加IP172.16.18.100Sep 25 17:32:25 node1 Keepalived_vrrp[16629]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth0 for 172.16.18.100###############备节点##########[root@node2 keepalived]# tail -20 /var/log/messagesSep 25 17:32:23 node2 Keepalived_healthcheckers[20357]: Interface queue is emptySep 25 17:32:23 node2 Keepalived_vrrp[20358]: Interface queue is emptySep 25 17:32:23 node2 Keepalived_healthcheckers[20357]: Netlink reflector reports IP 172.16.18.9 addedSep 25 17:32:23 node2 Keepalived_vrrp[20358]: Netlink reflector reports IP 172.16.18.9 addedSep 25 17:32:23 node2 Keepalived_healthcheckers[20357]: Netlink reflector reports IP fe80::20c:29ff:fe12:c8b5 addedSep 25 17:32:23 node2 Keepalived_vrrp[20358]: Netlink reflector reports IP fe80::20c:29ff:fe12:c8b5 addedSep 25 17:32:23 node2 Keepalived_vrrp[20358]: Registering Kernel netlink reflectorSep 25 17:32:23 node2 Keepalived_healthcheckers[20357]: Registering Kernel netlink reflectorSep 25 17:32:23 node2 Keepalived_healthcheckers[20357]: Registering Kernel netlink command channelSep 25 17:32:23 node2 Keepalived_vrrp[20358]: Registering Kernel netlink command channelSep 25 17:32:23 node2 Keepalived_vrrp[20358]: Registering gratuitous ARP shared channelSep 25 17:32:23 node2 Keepalived_healthcheckers[20357]: Opening file '/etc/keepalived/keepalived.conf'.Sep 25 17:32:23 node2 Keepalived_healthcheckers[20357]: Configuration is using : 6985 BytesSep 25 17:32:23 node2 Keepalived_vrrp[20358]: Opening file '/etc/keepalived/keepalived.conf'.Sep 25 17:32:23 node2 Keepalived_vrrp[20358]: Configuration is using : 62678 BytesSep 25 17:32:23 node2 Keepalived_vrrp[20358]: Using LinkWatch kernel netlink reflector...Sep 25 17:32:23 node2 Keepalived_vrrp[20358]: VRRP_Instance(VI_1) Entering BACKUP STATE #进入BACKUP状态Sep 25 17:32:23 node2 Keepalived_vrrp[20358]: VRRP sockpool: [ifindex(2), proto(112), fd(10,11)]Sep 25 17:32:23 node2 Keepalived_healthcheckers[20357]: Using LinkWatch kernel netlink reflector... |
测试：将node1关闭，node2会不会将地址取走？？
|
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
|
##############主节点############[root@node1 ~]# service keepalived stopStopping keepalived: [ OK ][root@node1 ~]# ip addr show1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:06:a6:49 brd ff:ff:ff:ff:ff:ff inet 172.16.18.7/16 brd 172.16.255.255 scope global eth0 inet6 fe80::20c:29ff:fe06:a649/64 scope link valid_lft forever preferred_lft forever###############备节点##########[root@node2 keepalived]# ip addr show1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:12:c8:b5 brd ff:ff:ff:ff:ff:ff inet 172.16.18.9/16 brd 172.16.255.255 scope global eth0 inet 172.16.18.100/32 scope global eth0 inet6 fe80::20c:29ff:fe12:c8b5/64 scope link valid_lft forever preferred_lft forever |
测试结果：这样是成立的，若node1重新上线会立即将VIP获取走。
如何使用keepalived调用外部脚本或手动执行命令实现VIP转移？？
思路：通过addr_script（脚本）定义检测机制；然后通过track_script在实例中追踪这个脚本。
如：下面这个检测机制
|
1
2
3
4
5
|
vrrp_script chk_mantaince_down { # chk_mantaince_down定义脚本的名称，可随意取 script "[[ -f /etc/keepalived/down ]] && exit 1 || exit 0" #命令（其实这里可以是自己定义好的脚本路径也可以是判断命令）#这里的意思是如果在这个文件下有down这个文件就表示期望这 个节点为备用状态。 interval 1 #每隔1秒钟执行一次 weight -2 #一旦命令执行失败，权重降低2个} |
（1）将此此检查机制应用到我们的示例中测试实现过程：
|
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
|
#############主节点###########[root@node1 keepalived]# vim keepalived.conf router_id LVS_DEVEL} vrrp_script chk_main { 脚本 script "[[ -f /etc/keepalived/down ]] && exit 1 || exit 0" interval 1 weight -2} vrrp_instance VI_1 { state MASTER interface eth0 virtual_router_id 157 #如果环境中操作者比较多，尽量在每次更改配置文件之后改变一下这个值，从而实现ARPs快速转接。 priority 100 advert_int 1 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 172.16.18.100 } track_script { #追踪脚本 chk_main }}##############备节点############[root@node2 keepalived]# vim keepalived.confsmtp_connect_timeout 30 router_id LVS_DEVEL} vrrp_script chk_main { script "[[ -f /etc/keepalived/down ]] && exit 1 || exit 0" interval 1 weight -2} vrrp_instance VI_1 { state BACKUP interface eth0 virtual_router_id 157 priority 99 advert_int 1 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 172.16.18.100 } track_script { chk_main }} |
（2）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
|
#######未添加文件之前：节点一########[root@node1 keepalived]# ip addr1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:06:a6:49 brd ff:ff:ff:ff:ff:ff inet 172.16.18.7/16 brd 172.16.255.255 scope global eth0 inet 172.16.18.100/32 scope global eth0 inet6 fe80::20c:29ff:fe06:a649/64 scope link valid_lft forever preferred_lft forever#######未添加文件之前：节点二########[root@node2 keepalived]# ip addr1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:12:c8:b5 brd ff:ff:ff:ff:ff:ff inet 172.16.18.9/16 brd 172.16.255.255 scope global eth0 inet6 fe80::20c:29ff:fe12:c8b5/64 scope link valid_lft forever preferred_lft forever############添加文件########[root@node1 keepalived]# touch down###########查看状态#########[root@node1 keepalived]# ip addr1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:06:a6:49 brd ff:ff:ff:ff:ff:ff inet 172.16.18.7/16 brd 172.16.255.255 scope global eth0 inet6 fe80::20c:29ff:fe06:a649/64 scope link valid_lft forever preferred_lft forever[root@node2 keepalived]# ip addr1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:12:c8:b5 brd ff:ff:ff:ff:ff:ff inet 172.16.18.9/16 brd 172.16.255.255 scope global eth0 inet 172.16.18.100/32 scope global eth0 inet6 fe80::20c:29ff:fe12:c8b5/64 scope link valid_lft forever preferred_lft forever |
如何在状态转换时进行通知？？
（1）keepalive内部提供了两个配置指令详细参考man keepalived.conf，一般在vrrp_instance或者vrrp_sync_group中使用中使用：
第一类指令：
|
1
2
3
4
5
6
|
# to MASTER transitionnotify_master /path/to_master.sh #转换为master状态时使用此脚本通知# to BACKUP transitionnotify_backup /path/to_backup.sh #转换为backup状态时使用此脚本通知# FAULT transitionnotify_fault "/path/fault.sh VG_1" #如果变成了fault就是用此脚本通知，如果脚本带有参数也就是有空格必须使用引号 |
第二类指令：使用notify直接引用
|
1
2
3
4
5
|
# $1 = "GROUP"|"INSTANCE" #参数1：必须能够指定接受组或实例# $2 = name of group or instance #这个组或实例的名称# $3 = target state of transition #指定转换成哪个状态进行通知的# ("MASTER"|"BACKUP"|"FAULT")notify /path/notify.sh notify #脚本的路径（自行写） |
（2）通知脚本定义：
|
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
|
[root@node1 keepalived]# vim notify.sh#!/bin/bash#vip=172.16.18.100 #指定VIPcontact='root@localhost' #通知给谁thisip=`ifconfig eth0 | awk '/inet addr:/{print $2}' | awk -F: '{print $2}'` #获取当前节点IP地址notify() { mailsubject="$thisip to be $1: $vip floating" mailbody="`date '+%F %H:%M:%S'`: vrrp transition, $thisip changed to be $1" echo $mailbody | mail -s "$mailsubject" $contact}case "$1" in master) notify master exit 0 ;; backup) notify backup exit 0 ;; fault) notify fault exit 0 ;; *) echo 'Usage: `basename $0` {master|backup|fault}' exit 1 ;;esac##########赋予此脚本执行权限###########[root@node1 keepalived]# chmod +x notify.sh###########测试脚本####################[root@node1 keepalived]# ./notify.sh master###########查看邮件信息################[root@node1 keepalived]# mailHeirloom Mail version 12.4 7/29/08. Type ? for help."/var/spool/mail/root": 1 message 1 new>N 1 root Wed Sep 25 22:24 18/693 "172.16.18.7 to be master: 172.16.18.100 flo"& 1 #第一封邮件Message 1:From root@node1.magedu.com Wed Sep 25 22:24:40 2013Return-Path: <root@node1.magedu.com>X-Original-To: root@localhostDelivered-To: root@localhost.magedu.comDate: Wed, 25 Sep 2013 22:24:39 +0800To: root@localhost.magedu.comSubject: 172.16.18.7 to be master: 172.16.18.100 floatingUser-Agent: Heirloom mailx 12.4 7/29/08Content-Type: text/plain; charset=us-asciiFrom: root@node1.magedu.com (root)Status: R2013-09-25 22:24:39: vrrp transition, 172.16.18.7 changed to be master #内容###########使用quit退出邮件############## |
|
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
|
#############配置测试状态转换##############[root@node1 keepalived]# vim keepalived.confvrrp_instance VI_1 { state MASTER interface eth0 virtual_router_id 157 priority 100 advert_int 1 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 172.16.18.100 } track_script { chk_main } notify_master "/etc/keepalived/notify.sh master" #指定切换到Master状态时执行的脚本 notify_backup "/etc/keepalived/notify.sh backup" #指定切换到Backup状态时执行的脚本 notify_fault "/etc/keepalived/notify.sh fault" #指定切换到Mfault状态时执行的脚本}##########注意将上面此代码写入备节点中############notify_master "/etc/keepalived/notify.sh master" #指定切换到Master状态时执行的脚本notify_backup "/etc/keepalived/notify.sh backup" #指定切换到Backup状态时执行的脚本notify_fault "/etc/keepalived/notify.sh fault" #指定切换到Mfault状态时执行的脚本##########脚本同样在备节点中存在##################[root@node1 keepalived]# scp notify.sh 172.16.18.9:/etc/keepalived/ |
（3）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
|
[root@node1 keepalived]# touch down[root@node1 keepalived]# ip addr1: 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 00:0c:29:06:a6:49 brd ff:ff:ff:ff:ff:ff inet 172.16.18.7/16 brd 172.16.255.255 scope global eth0 inet6 fe80::20c:29ff:fe06:a649/64 scope link valid_lft forever preferred_lft forever[root@node1 keepalived]# mail #节点1上Heirloom Mail version 12.4 7/29/08. Type ? for help."/var/spool/mail/root": 3 messages 2 unread 1 root Wed Sep 25 22:24 19/704 "172.16.18.7 to be master: 172.16.18.100 flo">U 2 root Wed Sep 25 22:47 19/703 "172.16.18.7 to be master: 172.16.18.100 flo" U 3 root Wed Sep 25 22:47 19/703 "172.16.18.7 to be backup: 172.16.18.100 flo"&[root@node2 keepalived]# mail #节点2上Heirloom Mail version 12.4 7/29/08. Type ? for help."/var/spool/mail/root": 3 messages 3 new>N 1 root Wed Sep 25 22:46 18/693 "172.16.18.9 to be backup: 172.16.18.100 flo" N 2 root Wed Sep 25 22:47 18/693 "172.16.18.9 to be backup: 172.16.18.100 flo" N 3 root Wed Sep 25 22:47 18/693 "172.16.18.9 to be master: 172.16.18.100 flo"& |
关于ipvs配置生成规则实现负载均衡和web服务器实现高可用等更多关于keepalived高级应用将在后续博客中持续更新，请继续关注！！谢谢！！
非抢占模式
master从故障中恢复后，不会抢占备份节点的vip
1> MASTER（192.168.1.201）：
global_defs {
router_id nginx_01 #标识本节点的名称，通常为hostname
}
vrrp_script chk_nginx {
script "/etc/keepalived/nginx_check.sh"
interval 2
weight -20
}
vrrp_instance VI_1 {
state BACKUP
interface enp0s3
virtual_router_id 51
mcast_src_ip 192.168.1.201
priority 100
advert_int 1
nopreempt
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.1.210
}
track_script {
chk_nginx # nginx存活状态检测脚本
}
}2> BACKUP（192.168.1.202）
global_defs {
router_id nginx_02
}
vrrp_script chk_nginx {
script "/etc/keepalived/nginx_check.sh"
interval 2
weight -20
}
vrrp_instance VI_1 {
state BACKUP
interface enp0s3
virtual_router_id 51
mcast_src_ip 192.168.1.202
priority 90
advert_int 1
nopreempt
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.1.210
}
track_script {
chk_nginx
}
}和非抢占模式的配置相比，只改了两个地方：
1> 在vrrp_instance块下两个节点各增加了nopreempt指令，表示不争抢vip
2> 节点的state都为BACKUP
两个keepalived节点都启动后，默认都是BACKUP状态，双方在发送组播信息后，会根据优先级来选举一个MASTER出来。由于两者都配置了nopreempt，所以MASTER从故障中恢复后，不会抢占vip。这样会避免VIP切换可能造成的服务延迟。