我尝试解决这个问题已经一整天了，但没有成功。
我有两台服务器，server1 和 server2，均运行 Ubuntu 14.04.5 LTS，并通过带 LACP 的 LAG 中继连接到 Cisco sg200-08 交换机。交换机 ip 为 172.128.1.254/24，服务器上的接口如下所示，包括相关 ip 的路由和 arp 表：
在server1上：
root@server1:~# ip addr show bond0
5: bond0: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default
link/ether 00:11:0a:10:03:29 brd ff:ff:ff:ff:ff:ff
inet 172.128.1.129/24 brd 172.128.1.255 scope global bond0
valid_lft forever preferred_lft forever
root@server1:~# ip addr show bond0.53
13: bond0.53@bond0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default
link/ether 00:11:0a:10:03:29 brd ff:ff:ff:ff:ff:ff
inet 192.168.53.1/24 brd 192.168.53.255 scope global bond0.53
valid_lft forever preferred_lft forever
root@server1:~# ip route get 192.168.53.2
192.168.53.2 dev bond0.53 src 192.168.53.1
cache
root@server1:~# arp -n | grep '192.168.53.2'
192.168.53.2 (incomplete) bond0.53
在server2上：
root@server2:~# ip addr show bond0
5: bond0: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default
link/ether 00:15:17:2e:ab:b4 brd ff:ff:ff:ff:ff:ff
inet 172.128.1.130/24 brd 172.128.1.255 scope global bond0
valid_lft forever preferred_lft foreve
root@server2:~# ip addr show bond0.53
22: bond0.53@bond0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default
link/ether 00:15:17:2e:ab:b4 brd ff:ff:ff:ff:ff:ff
inet 192.168.53.2/24 brd 192.168.53.255 scope global bond0.53
valid_lft forever preferred_lft forever
root@server2:~# ip route get 192.168.53.1
192.168.53.1 dev bond0.53 src 192.168.53.2
cache
root@server2:~# arp -n | grep '192.168.53.1'
192.168.53.1 ether 00:11:0a:10:03:29 C bond0.53
当我从服务器 1 ping 服务器 2 时，我看不到返回服务器 1 的 arp 回复：
root@server1:~# tcpdump -ennqt -i bond0 \( arp or icmp \)
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on bond0, link-type EN10MB (Ethernet), capture size 65535 bytes
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 28
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 28
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 28
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 28
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 28
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 28
但在 server2 端，我可以看到来自 server1 的 arp 请求和通过 VLAN53 发送回的答复：
root@server2:~# tcpdump -ennqt -i bond0 \( arp or icmp \)
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on bond0, link-type EN10MB (Ethernet), capture size 65535 bytes
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 64: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 46
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Reply 192.168.53.2 is-at 00:15:17:2e:ab:b4, length 28
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 64: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 46
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Reply 192.168.53.2 is-at 00:15:17:2e:ab:b4, length 28
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 64: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 46
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Reply 192.168.53.2 is-at 00:15:17:2e:ab:b4, length 28
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 64: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 46
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Reply 192.168.53.2 is-at 00:15:17:2e:ab:b4, length 28
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 64: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 46
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Reply 192.168.53.2 is-at 00:15:17:2e:ab:b4, length 28
00:11:0a:10:03:29 > ff:ff:ff:ff:ff:ff, 802.1Q, length 64: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.2 tell 192.168.53.1, length 46
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Reply 192.168.53.2 is-at 00:15:17:2e:ab:b4, length 28
对于相反方向的 ping，我只能在 server2 上看到这一点：
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 102: vlan 53, p 0, ethertype IPv4, 192.168.53.2 > 192.168.53.1: ICMP echo request, id 6506, seq 1, length 64
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 102: vlan 53, p 0, ethertype IPv4, 192.168.53.2 > 192.168.53.1: ICMP echo request, id 6506, seq 2, length 64
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 102: vlan 53, p 0, ethertype IPv4, 192.168.53.2 > 192.168.53.1: ICMP echo request, id 6506, seq 3, length 64
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 102: vlan 53, p 0, ethertype IPv4, 192.168.53.2 > 192.168.53.1: ICMP echo request, id 6506, seq 4, length 64
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 102: vlan 53, p 0, ethertype IPv4, 192.168.53.2 > 192.168.53.1: ICMP echo request, id 6506, seq 5, length 64
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.1 tell 192.168.53.2, length 28
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.1 tell 192.168.53.2, length 28
00:15:17:2e:ab:b4 > 00:11:0a:10:03:29, 802.1Q, length 46: vlan 53, p 0, ethertype ARP, Request who-has 192.168.53.1 tell 192.168.53.2, length 28
两侧均未设置防火墙、arptables 或 ebtables。内核 sysctl 未阻止 ICMP 流量。绑定已启动且运行正常。交换机在每个 LAG 中都有 2 个端口，配置为通向每个服务器的中继，并带有 vlan 1（本机/默认未标记）和 51、52、53、54 标记。我可以从交换机 ping 通 bond0 ip 172.128.1.129 和 172.128.1.130。我可以从连接到交换机的另一台 Linux PC（ip 为 172.128.1.5）ping 172.128.1.129（服务器 1），但不能 ping 172.128.1.130（服务器 2）。
提前感谢任何指点、想法和建议。
更正：我可以从网络上的第三个主机 ping 通这两台服务器
igorc@client:~$ ip -f inet addr show eth1
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
inet 172.128.1.5/24 brd 172.128.1.255 scope global dynamic eth1
valid_lft 22497sec preferred_lft 22497sec
igorc@client:~$ ping -c 2 172.128.1.129
PING 172.128.1.129 (172.128.1.129) 56(84) bytes of data.
64 bytes from 172.128.1.129: icmp_seq=1 ttl=64 time=0.618 ms
64 bytes from 172.128.1.129: icmp_seq=2 ttl=64 time=0.541 ms
--- 172.128.1.129 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1000ms
rtt min/avg/max/mdev = 0.541/0.579/0.618/0.045 ms
igorc@client:~$ ping -c 2 172.128.1.130
PING 172.128.1.130 (172.128.1.130) 56(84) bytes of data.
64 bytes from 172.128.1.130: icmp_seq=1 ttl=64 time=0.645 ms
64 bytes from 172.128.1.130: icmp_seq=2 ttl=64 time=0.693 ms
--- 172.128.1.130 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 0.645/0.669/0.693/0.024 ms
更新：两台服务器上的绑定
root@server1:~# cat /proc/net/bonding/bond0
Ethernet Channel Bonding Driver: v3.7.1 (April 27, 2011)
Bonding Mode: IEEE 802.3ad Dynamic link aggregation
Transmit Hash Policy: layer2 (0)
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 100
Down Delay (ms): 100
802.3ad info
LACP rate: slow
Min links: 0
Aggregator selection policy (ad_select): stable
Active Aggregator Info:
Aggregator ID: 1
Number of ports: 1
Actor Key: 17
Partner Key: 1
Partner Mac Address: 00:00:00:00:00:00
Slave Interface: eth2
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 2
Permanent HW addr: 00:11:0a:10:03:29
Aggregator ID: 1
Slave queue ID: 0
Slave Interface: eth0
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 2
Permanent HW addr: 00:11:0a:10:03:28
Aggregator ID: 2
Slave queue ID: 0
root@server2:~# cat /proc/net/bonding/bond0
Ethernet Channel Bonding Driver: v3.7.1 (April 27, 2011)
Bonding Mode: IEEE 802.3ad Dynamic link aggregation
Transmit Hash Policy: layer2 (0)
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 100
Down Delay (ms): 100
802.3ad info
LACP rate: slow
Min links: 0
Aggregator selection policy (ad_select): stable
Active Aggregator Info:
Aggregator ID: 2
Number of ports: 1
Actor Key: 17
Partner Key: 1
Partner Mac Address: 00:00:00:00:00:00
Slave Interface: p1p1
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:15:17:2e:ab:b4
Aggregator ID: 1
Slave queue ID: 0
Slave Interface: p1p2
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:15:17:2e:ab:b5
Aggregator ID: 2
Slave queue ID: 0
答案1
解决了。我错误地将 Cisco 交换机中的 LAG 设置为静态而不是动态，这阻止了 LACP 的使用。嵌入的图像不会显示，可能是因为我的帐户中缺少积分，但无论如何都会附加它。
现在一切看起来好多了：
root@server1:~# cat /proc/net/bonding/bond0
Ethernet Channel Bonding Driver: v3.7.1 (April 27, 2011)
Bonding Mode: IEEE 802.3ad Dynamic link aggregation
Transmit Hash Policy: layer2+3 (2)
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 100
Down Delay (ms): 100
802.3ad info
LACP rate: fast
Min links: 0
Aggregator selection policy (ad_select): stable
Active Aggregator Info:
Aggregator ID: 1
**Number of ports: 2**
Actor Key: 17
Partner Key: 10
**Partner Mac Address: 20:bb:c0:78:7e:9b**
Slave Interface: eth0
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:11:0a:10:03:28
**Aggregator ID: 1**
Slave queue ID: 0
Slave Interface: eth2
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:11:0a:10:03:29
**Aggregator ID: 1**
Slave queue ID: 0
更改以粗体突出显示（如果在代码小部件中可见），首先，端口数正确设置为 2 而不是之前的 1，然后，聚合器 ID 现在正确地具有两个从属设备的相同值，最后，合作伙伴 Mac 地址现在具有一个值（与之前的 00:00:00:00:00:00 相比）表示对等体之间交换 LACP UDP 消息。