今天我获得了一个有2个IP地址的VPS。
eth0当我打开IP时ping -I eth0 www.google.com，数据包丢失率为 0％，但是当我打开 IP 时，ping -I eth1 www.goole.com数据包丢失率为 100％。
这是ifconfig输出：
eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500
inet 185.8.49.12 netmask 255.255.255.0 broadcast 185.8.49.255
inet6 fe80::250:56ff:fe84:5ed6 prefixlen 64 scopeid 0x20<link>
ether 00:50:56:84:5e:d6 txqueuelen 1000 (Ethernet)
RX packets 5716 bytes 398892 (389.5 KiB)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 933 bytes 294738 (287.8 KiB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
eth1: flags=4099<UP,BROADCAST,MULTICAST> mtu 1500
inet 185.8.49.157 netmask 255.255.255.0 broadcast 185.8.49.255
ether 00:50:56:84:5e:d7 txqueuelen 1000 (Ethernet)
RX packets 0 bytes 0 (0.0 B)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 0 bytes 0 (0.0 B)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
lo: flags=73<UP,LOOPBACK,RUNNING> mtu 65536
inet 127.0.0.1 netmask 255.0.0.0
inet6 ::1 prefixlen 128 scopeid 0x10<host>
loop txqueuelen 0 (Local Loopback)
RX packets 56 bytes 8896 (8.6 KiB)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 56 bytes 8896 (8.6 KiB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
这是输出ifcfg-eth0：
DEVICE=eth0
NM_CONTROLLED=no
ONBOOT=yes
TYPE=Ethernet
DEFROUTE=yes
IPV4_FAILURE_FATAL=yes
IPV6INIT=no
NAME="System eth0"
BOOTPROTO=none
IPADDR=185.8.49.12
NETMASK=255.255.255.0
GATEWAY=185.8.49.1
DNS1=62.149.128.4
DNS2=62.149.132.4
这是输出ifcfg-eth1：
DEVICE=eth1
NM_CONTROLLED=no
ONBOOT=yes
TYPE=Ethernet
DEFROUTE=yes
IPV4_FAILURE_FATAL=yes
IPV6INIT=no
NAME="System eth1"
BOOTPROTO=none
IPADDR=185.8.49.157
NETMASK=255.255.255.0
GATEWAY=185.8.49.1
我尝试重新启动系统，但毫无作用。
答案1
两个接口配置上的“DEFROUTE=yes”并没有实现您所想的效果。
重启（清除你所做的任何修改）并运行“ip route”。
你应该看到类似这样的内容：
# ip route
default via 185.8.49.1 dev eth0
185.8.49.0/24 dev eth0 proto kernel scope link src 185.8.49.12
185.8.49.0/24 dev eth1 proto kernel scope link src 185.8.49.157
当您发出“ping -I eth1 8.8.8.8”时，由于系统未配置可通过 eth1 访问的默认网关，因此 ARP 请求会从所有接口发送出去，以在本地网络上查找 8.8.8.8：
# ping -I eth1 8.8.8.8
PING 8.8.8.8 (8.8.8.8) from 185.8.49.157 eth1: 56(84) bytes of data.
From 185.8.49.157 icmp_seq=1 Destination Host Unreachable
From 185.8.49.157 icmp_seq=2 Destination Host Unreachable
From 185.8.49.157 icmp_seq=3 Destination Host Unreachable
From 185.8.49.157 icmp_seq=4 Destination Host Unreachable
# tcpdump -ni eth0 'arp'
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth0, link-type EN10MB (Ethernet), capture size 65535 bytes
05:07:42.821526 ARP, Request who-has 8.8.8.8 tell 185.8.49.157, length 46
05:07:43.821185 ARP, Request who-has 8.8.8.8 tell 185.8.49.157, length 46
05:07:44.823000 ARP, Request who-has 8.8.8.8 tell 185.8.49.157, length 46
# tcpdump -ni eth1 'arp'
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth1, link-type EN10MB (Ethernet), capture size 65535 bytes
05:07:42.820834 ARP, Request who-has 8.8.8.8 tell 185.8.49.157, length 28
05:07:43.820864 ARP, Request who-has 8.8.8.8 tell 185.8.49.157, length 28
05:07:44.822841 ARP, Request who-has 8.8.8.8 tell 185.8.49.157, length 28
（显然 Google 的 DNS 服务器与您的 VPS 不在同一子网中。）
继续尝试添加第二条默认路由：
# ip route add default via 185.8.49.1 dev eth1
RTNETLINK answers: File exists
看起来系统不会轻易接受多个默认路由。
这很有道理——否则设备如何知道通过多个网关中的哪一个发送数据包？它会向每个网关发送一份副本...然后处理多个返回数据包吗？还是它会以不确定的方式任意发送数据包（这是一个难以排除故障的噩梦）？
大概它可以进行负载平衡，所以让我们试试：
#ip route delete default
#ip route add default scope global nexthop via 185.8.49.1 dev eth0 weight 1 nexthop via 185.8.49.1 dev eth1 weight 1
#ip ro
default
nexthop via 185.8.49.1 dev eth0 weight 1
nexthop via 185.8.49.1 dev eth1 weight 1
...
但它有效吗？
# ping 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=54 time=17.6 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=54 time=18.7 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=54 time=17.1 ms
64 bytes from 8.8.8.8: icmp_seq=4 ttl=54 time=15.3 ms
^C
--- 8.8.8.8 ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3005ms
rtt min/avg/max/mdev = 15.337/17.227/18.762/1.241 ms
# tcpdump -ni eth0 icmp
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth0, link-type EN10MB (Ethernet), capture size 65535 bytes
05:46:31.837933 IP 185.8.49.12 > 8.8.8.8: ICMP echo request, id 2382, seq 1, length 64
05:46:31.855566 IP 8.8.8.8 > 185.8.49.12: ICMP echo reply, id 2382, seq 1, length 64
05:46:33.842373 IP 185.8.49.12 > 8.8.8.8: ICMP echo request, id 2382, seq 3, length 64
05:46:33.859469 IP 8.8.8.8 > 185.8.49.12: ICMP echo reply, id 2382, seq 3, length 64
# tcpdump -ni eth1 icmp
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth1, link-type EN10MB (Ethernet), capture size 65535 bytes
05:46:32.840535 IP 185.8.49.157 > 8.8.8.8: ICMP echo request, id 2382, seq 2, length 64
05:46:32.859029 IP 8.8.8.8 > 185.8.49.157: ICMP echo reply, id 2382, seq 2, length 64
05:46:34.843725 IP 185.8.49.157 > 8.8.8.8: ICMP echo request, id 2382, seq 4, length 64
05:46:34.859020 IP 8.8.8.8 > 185.8.49.157: ICMP echo reply, id 2382, seq 4, length 64
好了！
现在由你来决定是否要进行负载平衡真的需要并愿意提供支持。