一、简介
二、模式
注*建议学之前先看这个文章
http://blog.csdn.net/galdys/article/details/7506495
1、balance-rr
轮转策略：按照顺序轮流使用每个接口来发送和接收数据包，提高了负载均衡的能力，但是需要交换机的支持，因为该模式所有端口的mac地址相同，交换机不能识别多个一个Mac地址对应多个端口，所以需要做链路聚合 ----Ethernet Channel
mode=0
2、active-backup
主备策略：只使用其中一个网卡，当网卡损坏时激活另外的一个网卡
mode=1
3、balance-xor
平衡策略： 基于所选择的传送hash策略
mode=2
4、
广播策略：所有卡上都接收和发送报文，提高了容错性
mode=3
5、802.3ad
链路聚合模式
802.3ad策略： LACP 链路聚合模式，交换机得开启链路聚合，将两张网卡当做一张来使用，提高链路性能，需要交换机支持802.3ad 动态链路聚合和服务器硬件支持ethtool。
mode=4
6、balance-tlb
自适应传输负载均衡
不需要配置交换机。在每个网卡会根据上根据当前的负载情况来分配流量。如果一个网卡出现故障，那么另外一个网卡会接管故障网卡
mode=5
7、balance-alb
自适应负载均衡：包括balance-tlb，不需要另外配置交换机，接受负载均衡通过ARP协商实现
当本机发送ARP请求时，bonding驱动把对端的IP信息从ARP包中复制并保存下来，当ARP应答从对端到达 时，bonding驱动把它的硬件地址提取出来，并发起一个ARP应答给bond中的某个slave。使用ARP协商进行负载均衡的一个问题是：每次广播 ARP请求时都会使用bond的硬件地址，因此对端学习到这个硬件地址后，接收流量将会全部流向当前的slave。这个问题可以通过给所有的对端发送更新 （ARP应答）来解决，应答中包含他们独一无二的硬件地址，从而导致流量重新分布。当新的slave加入到bond中时，或者某个未激活的slave重新 激活时，接收流量也要重新分布。接收的负载被顺序地分布（round robin）在bond中最高速的slave上
当某个链路被重新接上，或者一个新的slave加入到bond中，接收流量在所有当前激活的slave中全部重新分配，通过使用指定的MAC地址给每个 client发起ARP应答。下面介绍的updelay参数必须被设置为某个大于等于switch(交换机)转发延时的值，从而保证发往对端的ARP应答 不会被switch(交换机)阻截。
mode=6
三、安装
准备
多网卡主机一台
网卡桥接
选择局域网内空闲ip地址一枚
[root@localhost ~]# ip a
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
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eno16777736: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:fd:1d:cf brd ff:ff:ff:ff:ff:ff
inet 192.168.3.151/24 brd 192.168.3.255 scope global dynamic eno16777736
valid_lft 258897sec preferred_lft 258897sec
inet6 fe80::20c:29ff:fefd:1dcf/64 scope link
valid_lft forever preferred_lft forever
3: eno33554984: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:fd:1d:d9 brd ff:ff:ff:ff:ff:ff
inet 192.168.3.154/24 brd 192.168.3.255 scope global dynamic eno33554984
valid_lft 258897sec preferred_lft 258897sec
inet6 fe80::20c:29ff:fefd:1dd9/64 scope link
valid_lft forever preferred_lft forever
4: eno50332208: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:fd:1d:e3 brd ff:ff:ff:ff:ff:ff
inet 192.168.3.155/24 brd 192.168.3.255 scope global dynamic eno50332208
valid_lft 258897sec preferred_lft 258897sec
inet6 fe80::20c:29ff:fefd:1de3/64 scope link
valid_lft forever preferred_lft forever
[root@localhost ~]#
#查看硬件状态
[root@localhost ~]# nmcli device status
设备 类型 状态 CONNECTION
eno16777736 ethernet 连接的 eno16777736
eno33554984 ethernet 连接的 有线连接 1
eno50332208 ethernet 连接的 有线连接 2
lo loopback 未管理 --
#查看linux是否支持Bond
[root@localhost ~]# modinfo bonding
filename: /lib/modules/3.10.0-229.el7.x86_64/kernel/drivers/net/bonding/bonding.ko
alias: rtnl-link-bond
author: Thomas Davis, tadavis@lbl.gov and many others
description: Ethernet Channel Bonding Driver, v3.7.1
version: 3.7.1
license: GPL
..........
..........
#以上信息反映机器支持bond
#本人选取其中两种方式来制作bond绑定，一种是active-backup，另外一种选择balance-tlb，并且做相应测试
#方式1:nmcli方式
使用nmcli命令来创建bond卡
nmcli connection add type bond ifname bond1 mode 1
nmcli connection add type bond-slave ifname eno33554984 master bond1
nmcli connection add type bond-slave ifname eno50332208 master bond1
nmcli connection add type bond-slave ifname eno16777736 master bond1
#connection add 添加新的连接
#con-name 连接名称 暂时没使用
#type 设备类型
#ifname 接口设备名称
#master 指定该硬件的master
# 执行完成以上命令后会生成以下文件
[root@localhost ~]# cd /etc/sysconfig/network-scripts/
[root@localhost network-scripts]# ll
总用量 248
-rw-r--r--. 1 root root 328 3月 29 18:08 ifcfg-bond-bond1
-rw-r--r--. 1 root root 137 3月 29 18:08 ifcfg-bond-slave-eno16777736
-rw-r--r--. 1 root root 137 3月 29 18:08 ifcfg-bond-slave-eno33554984
-rw-r--r--. 1 root root 137 3月 29 18:08 ifcfg-bond-slave-eno50332208
#修改已经生成的配置文件
[root@localhost network-scripts]# vim ifcfg-bond-bond1
DEVICE=bond1
BONDING_OPTS=mode=active-backup
TYPE=Bond
BONDING_MASTER=yes
BOOTPROTO=static
DEFROUTE=yes
PEERDNS=yes
PEERROUTES=yes
IPV4_FAILURE_FATAL=no
IPV6INIT=yes
IPV6_AUTOCONF=yes
IPV6_DEFROUTE=yes
IPV6_PEERDNS=yes
IPV6_PEERROUTES=yes
IPV6_FAILURE_FATAL=no
NAME=bond-bond1
UUID=5ed04865-5802-4fd5-9d80-917d1fadae70
ONBOOT=yes
IPADDR=192.168.3.200
PREFIX=24
GATEWAY=192.168.3.254
DNS1=192.168.30.202
#重启网卡
[root@localhost network-scripts]# systemctl restart network
#查看网络和路由状态
[root@localhost ~]# ip a
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
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eno16777736: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master bond1 state UP qlen 1000
link/ether 00:0c:29:fd:1d:cf brd ff:ff:ff:ff:ff:ff
3: eno33554984: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master bond1 state UP qlen 1000
link/ether 00:0c:29:fd:1d:cf brd ff:ff:ff:ff:ff:ff
4: eno50332208: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master bond1 state UP qlen 1000
link/ether 00:0c:29:fd:1d:cf brd ff:ff:ff:ff:ff:ff
6: bond1: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP
link/ether 00:0c:29:fd:1d:cf brd ff:ff:ff:ff:ff:ff
inet 192.168.3.200/24 brd 192.168.3.255 scope global bond1
valid_lft forever preferred_lft forever
inet6 fe80::20c:29ff:fefd:1dcf/64 scope link
valid_lft forever preferred_lft forever
[root@localhost ~]# ip route
default via 192.168.3.254 dev bond1 proto static metric 300
192.168.3.0/24 dev bond1 proto kernel scope link src 192.168.3.200 metric 300
[root@localhost ~]# cat /proc/net/bonding/bond1
Ethernet Channel Bonding Driver: v3.7.1 (April 27, 2011)
Bonding Mode: fault-tolerance (active-backup)
Primary Slave: None
Currently Active Slave: eno16777736
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0
Slave Interface: eno16777736
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:0c:29:fd:1d:cf
Slave queue ID: 0
Slave Interface: eno33554984
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:0c:29:fd:1d:d9
Slave queue ID: 0
Slave Interface: eno50332208
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:0c:29:fd:1d:e3
Slave queue ID: 0
#标红为当前打开的网卡
#测试
#三张网卡随意关闭任意两张
#关闭掉两张网卡后，会有一次断点，切换网卡的过程
#网卡已经切换为eno50332208
[root@localhost ~]# cat /proc/net/bonding/bond1
Ethernet Channel Bonding Driver: v3.7.1 (April 27, 2011)
Bonding Mode: fault-tolerance (active-backup)
Primary Slave: None
Currently Active Slave: eno50332208
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0
Slave Interface: eno16777736
MII Status: down
Speed: Unknown
Duplex: Unknown
Link Failure Count: 1
Permanent HW addr: 00:0c:29:fd:1d:cf
Slave queue ID: 0
Slave Interface: eno33554984
MII Status: down
Speed: Unknown
Duplex: Unknown
Link Failure Count: 1
Permanent HW addr: 00:0c:29:fd:1d:d9
Slave queue ID: 0
Slave Interface: eno50332208
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:0c:29:fd:1d:e3
Slave queue ID: 0
#实验完成
#删除原有配置
#清除bond网卡配置
[root@localhost ~]# cd /etc/sysconfig/network-scripts/
[root@localhost network-scripts]# rm -rf ifcfg-bond-*
#显示所有活动的连接
[root@localhost network-scripts]# nmcli connection show
名称 UUID 类型 设备
有线连接 2 4013b28a-31dd-43b0-9b32-6582a20353f9 802-3-ethernet --
有线连接 1 9cca4859-5085-4cb3-a660-4664fac689c5 802-3-ethernet --
eno16777736 5dcd1c25-9d8b-4c0f-967c-cd0b250ff929 802-3-ethernet --
bond-slave-eno16777736 24c73fe2-9c55-4202-92f3-d4ff259a6d84 802-3-ethernet eno16777736
bond-slave-eno50332208 ea727513-602e-450e-ad36-dd15cc5d2e82 802-3-ethernet eno50332208
bond-slave-eno33554984 ca74b313-4a4e-44e3-bb5b-84d5dfa1d3aa 802-3-ethernet eno33554984
bond-bond1 5ed04865-5802-4fd5-9d80-917d1fadae70 bond bond1
#删除网卡配置连接
[root@localhost network-scripts]# nmcli con del bond-slave-eno50332208
[root@localhost network-scripts]# nmcli con del bond-slave-eno33554984
[root@localhost network-scripts]# cat /proc/net/bonding/bond1
Ethernet Channel Bonding Driver: v3.7.1 (April 27, 2011)
Bonding Mode: fault-tolerance (active-backup)
Primary Slave: None
Currently Active Slave: eno16777736
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0
Slave Interface: eno16777736
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 1
Permanent HW addr: 00:0c:29:fd:1d:cf
Slave queue ID: 0
[root@localhost network-scripts]# nmcli con del bond-slave-eno16777736
[root@localhost network-scripts]# nmcli con del bond-bond1
网卡恢复
#方式2:cli方式
以文件方式配置Bond
#备份原有网卡配置文件
[root@localhost ~]# mkdir -p /data/backup/network
[root@localhost ~]# cd /etc/sysconfig/network-scripts/
[root@localhost ~]# cp ifcfg-eno* /data/backup/network/
#因为是新加入的卡,所以没有配置文件，但是硬件信息中是可以查到的
[root@localhost network-scripts]# nmcli device status
设备 类型 状态 CONNECTION
eno16777736 ethernet 连接的 eno16777736
eno33554984 ethernet 连接的 有线连接 1
eno50332208 ethernet 连接的 有线连接 2
lo loopback 未管理 --
[root@localhost network-scripts]# ll
总用量 236
-rw-r--r--. 1 root root 295 8月 25 2015 ifcfg-eno16777736
-rw-r--r--. 1 root root 295 3月 29 19:13 ifcfg-eno33554984
[root@localhost network-scripts]# cp ifcfg-eno16777736 ifcfg-eno33554984
[root@localhost network-scripts]# vim ifcfg-bond5
DEVICE=bond5
BONDING_OPTS="mode=balance-tlb"
TYPE=Bond
BONDING_MASTER=yes
BOOTPROTO=static
DEFROUTE=yes
PEERDNS=yes
PEERROUTES=yes
IPV4_FAILURE_FATAL=no
NAME=bond5
ONBOOT=yes
IPADDR=192.168.3.200
PREFIX=24
GATEWAY=192.168.3.254
DNS1=192.168.30.202
[root@localhost network-scripts]# vim ifcfg-eno16777736
TYPE=Ethernet
#BOOTPROTO=dhcp
DEFROUTE=yes
PEERDNS=yes
PEERROUTES=yes
IPV4_FAILURE_FATAL=yes
NAME=eno16777736
DEVICE=eno16777736
ONBOOT=yes
MASTER=bond5
SLAVE=yes
[root@localhost network-scripts]# vim ifcfg-eno33554984
TYPE=Ethernet
#BOOTPROTO=dhcp
DEFROUTE=yes
PEERDNS=yes
PEERROUTES=yes
IPV4_FAILURE_FATAL=yes
NAME=eno33554984
DEVICE=eno33554984
ONBOOT=yes
MASTER=bond5
SLAVE=yes
[root@localhost network-scripts]# vim ifcfg-eno50332208
TYPE=Ethernet
#BOOTPROTO=dhcp
DEFROUTE=yes
PEERDNS=yes
PEERROUTES=yes
IPV4_FAILURE_FATAL=yes
NAME=eno50332208
DEVICE=eno50332208
ONBOOT=yes
MASTER=bond5
SLAVE=yes
#重启网卡
[root@localhost network-scripts]# systemctl restart network
[root@localhost ~]# nmcli connection show
名称 UUID 类型 设备
eno50332208 d2665055-8e83-58f1-e9e3-49a5fb133641 802-3-ethernet eno50332208
eno16777736 13756690-ac77-b776-4fc1-f5535cee6f16 802-3-ethernet eno16777736
有线连接 2 4013b28a-31dd-43b0-9b32-6582a20353f9 802-3-ethernet --
有线连接 1 9cca4859-5085-4cb3-a660-4664fac689c5 802-3-ethernet --
eno33554984 fb67dbad-ec81-39b4-42b1-ebf975c3ff13 802-3-ethernet eno33554984
bond5 d1a49a62-2f2d-2e41-a63d-be3a81b6a2fb bond bond5
[root@localhost ~]# ip a
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
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eno16777736: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master bond5 state UP qlen 1000
link/ether 00:0c:29:fd:1d:cf brd ff:ff:ff:ff:ff:ff
3: eno33554984: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master bond5 state UP qlen 1000
link/ether 00:0c:29:fd:1d:d9 brd ff:ff:ff:ff:ff:ff
4: eno50332208: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master bond5 state UP qlen 1000
link/ether 00:0c:29:fd:1d:e3 brd ff:ff:ff:ff:ff:ff
7: bond5: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP
link/ether 00:0c:29:fd:1d:cf brd ff:ff:ff:ff:ff:ff
inet 192.168.3.200/24 brd 192.168.3.255 scope global bond5
valid_lft forever preferred_lft forever
inet6 fe80::a81d:6bff:feac:5cd6/64 scope link
valid_lft forever preferred_lft forever
#个人建议各位跟balance-rr 模式做个对比
[root@localhost ~]# cat /proc/net/bonding/bond5
Ethernet Channel Bonding Driver: v3.7.1 (April 27, 2011)
Bonding Mode: transmit load balancing
Primary Slave: None
Currently Active Slave: eno16777736
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0
Slave Interface: eno16777736
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:0c:29:fd:1d:cf
Slave queue ID: 0
Slave Interface: eno33554984
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:0c:29:fd:1d:d9
Slave queue ID: 0
Slave Interface: eno50332208
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:0c:29:fd:1d:e3
Slave queue ID: 0
注*模式选择有以下这些
balance-rr(0)
active-backup(1)
balance-xor (2)
broadcast (3)
802.3ad (4)
balance-tlb (5)
balance-alb (6)
#暂时就这么多，希望各位多多指点！！
本文出自 “张浩宇” 博客，请务必保留此出处http://zhanghaoyu.blog.51cto.com/8202592/1911345
原文地址：http://zhanghaoyu.blog.51cto.com/8202592/1911345
