容器云负载均衡之五：使用keepalived管理基于容器的IPVS

一、前言

IPVS的director上会绑定一个VIP，这个VIP是下游客户的唯一接入点，所以这个VIP需要满足高可用性的要求。另外，不同的VIP策略需要根据后端real server的部署状态进行更新。

Keepalived是一个开源软件，主要提供loadbalancing（负载均衡）和 high-availability（高可用）功能，负载均衡实现需要依赖Linux的虚拟服务内核模块（IPVS），而高可用是通过VRRP协议实现多台机器之间的故障转移服务。 

转载自https://blog.csdn.net/cloudvtech

二、配置keepalived

**2.1 测试环境配置 **

1
2
3
4
5
6
1Director node 1: 200.222.0.73

2Director node 2: 200.222.0.74

3Real Server 1: 200.222.0.87

4Real Server 2: 200.222.0.89

5VIP: 200.222.0.113 

6

2.2 设置iptables允许VRRP协议交互

1
2
1iptables -t filter -A IN_public_allow -p udp -m udp --dport 112  -m conntrack --ctstate NEW -j ACCEPT

2

2.3 在CentOS容器内安装keepalived

1
2
3
4
5
6
1docker pull centos

2docker run -td --privileged --net=host --name=keepalived centos

3docker exec -it keepalived bash

4yum install -y net-tools iproute

5yum install -y keepalived ipvsadm

6

将容器打包成docker image

2.4 在两个Director node启动容器

2.5 配置Director node 1的keepalived

/etc/keepalived/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
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
1global_defs {

2   notification_email {

3     sysadmin@mydomain.com

4     support@mydomain.com

5   }

6   notification_email_from lb1@mydomain.com

7   smtp_server localhost

8   smtp_connect_timeout 30

9}

10

11vrrp_instance VI_1 {

12    state MASTER

13    interface ens192

14    virtual_router_id 51

15    priority 100

16    advert_int 1

17    unicast_src_ip 200.222.0.73

18    unicast_peer {

19        200.222.0.74

20    }

21    authentication {

22        auth_type PASS

23        auth_pass 1111

24    }

25    virtual_ipaddress {

26        200.222.0.113

27    }

28}

29

30virtual_server 200.222.0.113 80 {

31    delay_loop 6

32    lb_algo wlc

33    lb_kind DR

34    persistence_timeout 600

35    protocol TCP

36

37    real_server 200.222.0.87 80 {

38        weight 100

39        TCP_CHECK {

40            connect_timeout 3

41        }

42    }

43    real_server 200.222.0.89 80 {

44        weight 100

45        TCP_CHECK {

46            connect_timeout 3

47        }

48    }

49}

50

2.6 配置Director node 2的keepalived

/etc/keepalived/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
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
1global_defs {

2   notification_email {

3     sysadmin@mydomain.com

4     support@mydomain.com

5   }

6   notification_email_from lb1@mydomain.com

7   smtp_server localhost

8   smtp_connect_timeout 30

9}

10

11vrrp_instance VI_1 {

12    state BACKUP

13    interface ens192

14    virtual_router_id 51

15    priority 99

16    advert_int 1

17    unicast_src_ip 200.222.0.74

18    unicast_peer {

19        200.222.0.73

20    }

21    authentication {

22        auth_type PASS

23        auth_pass 1111

24    }

25    virtual_ipaddress {

26        200.222.0.113

27    }

28}

29

30virtual_server 200.222.0.113 80 {

31    delay_loop 6

32    lb_algo wlc

33    lb_kind DR

34    persistence_timeout 600

35    protocol TCP

36

37    real_server 200.222.0.87 80 {

38        weight 100

39        TCP_CHECK {

40            connect_timeout 3

41        }

42    }

43    real_server 200.222.0.89 80 {

44        weight 100

45        TCP_CHECK {

46            connect_timeout 3

47        }

48    }

49}

50

2.7 启动Director node 1/2的keepalived

可以看到keepalived仅仅在node1上绑定了VIP

2.8 查看Director node 1/2的VRRP协议交互信息

tcpdump -vvv -an -i ens192 | grep "vrid 51" 

可以看到active node持续向backup node发送VRRP协议信息

2.9 使用keepalived进行IPVS director failover

在Director node1运行pkill keepalived 

在Director node1可以看到VIP被绑定

也可以看到现在是新的active node(Director node 2)持续向node1发送VRRP协议信息

2.10 重新启动node1的keepalived，node1继续成为active的IPVS director

转载自https://blog.csdn.net/cloudvtech