記一次bond引發的網絡故障

本案中3個關鍵服務器
物理服務器：192.168.6.63，簡稱P，(Physical server)
KVM-VM：192.168.6.150，是物理服務器P上的一個KVM虛機，簡稱VM
NAS：外部NAS服務器，用來作ping/arp測試服務器，簡稱NAS

物理服務器P的配置：

#uname -a
Linux cz63 4.15.18-11-pve #1 SMP PVE 4.15.18-34 (Mon, 25 Feb 2019 14:51:06 +0100) x86_64 GNU/Linux

#cat /etc/network/interfaces
auto lo
iface lo inet loopback

auto enp3s0f0
iface enp3s0f0 inet manual

auto enp3s0f1
iface enp3s0f1 inet manual

auto ens1f0
iface ens1f0 inet manual

auto ens1f1
iface ens1f1 inet manual

auto bond0
iface bond0 inet manual
    bond-slaves enp3s0f0 enp3s0f1
    bond-miimon 100
    bond-mode balance-rr

auto bond1
iface bond1 inet manual
    bond-slaves ens1f0 ens1f1
    bond-miimon 100
    bond-mode balance-rr

auto vmbr0
iface vmbr0 inet static
    address  192.168.6.63
    netmask  255.255.255.0
    gateway  192.168.6.1
    bridge-ports bond0
    bridge-stp off
    bridge-fd 0

auto vmbr1
iface vmbr1 inet static
    address  10.1.1.63
    netmask  255.255.255.0
    bridge-ports bond1
    bridge-stp off
    bridge-fd 0
#brctl show
bridge name bridge id       STP enabled interfaces
vmbr0       8000.ac1f6b342094   no      bond0
                            tap401000001i0
vmbr1       8000.74a4b500e768   no      bond1

安裝KVM-VM，以後VM與外部服務器的鏈接很是不穩定，90%狀況下會出現ping不可達。這裏測試了centos、ubuntu、win7都是同樣的不穩定。
在VM內執行ping 192.168.6.40（NAS）
結果ping不通。
此時，在P上執行
tcpdump -leni vmbr0 arp
tcpdump -leni tap401000001i0 arp
發現狀況以下：

P上
#tcpdump -leni vmbr0 arp | grep 2a:f0:5f:ae:c9:8b
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 46
00:50:56:87:86:b9 > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.40 is-at 00:50:56:87:86:b9, length 46
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 46
00:50:56:87:86:b9 > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.40 is-at 00:50:56:87:86:b9, length 46
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 46
00:50:56:87:86:b9 > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.40 is-at 00:50:56:87:86:b9, length 46
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 46
00:50:56:87:86:b9 > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.40 is-at 00:50:56:87:86:b9, length 46
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 46
00:50:56:87:86:b9 > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.40 is-at 00:50:56:87:86:b9, length 46

P上
#tcpdump -leni tap401000001i0  arp | grep 2a:f0:5f:ae:c9:8b
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 46
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 46
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 46
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 46
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 46
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 46
2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 42: Request who-has 192.168.6.40 (ff:ff:ff:ff:ff:ff) tell 192.168.6.150, length 28

NAS上
# tcpdump -leni vmx0 arp | grep c9:8b
02:03:50.930907 2a:f0:5f:ae:c9:8b > 00:50:56:87:86:b9, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.40 tell 192.168.6.150, length 46
02:03:50.930923 00:50:56:87:86:b9 > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 42: Reply 192.168.6.40 is-at 00:50:56:87:86:b9, length 28
02:04:02.669823 2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.32 tell 192.168.6.150, length 46
02:04:02.670131 00:50:56:87:e3:6d > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.32 is-at 00:50:56:87:e3:6d, length 46
02:04:03.670770 2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.32 tell 192.168.6.150, length 46
02:04:03.671059 00:50:56:87:e3:6d > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.32 is-at 00:50:56:87:e3:6d, length 46
02:04:04.672736 2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.32 tell 192.168.6.150, length 46
02:04:04.672992 00:50:56:87:e3:6d > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.32 is-at 00:50:56:87:e3:6d, length 46
02:04:06.671878 2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.32 tell 192.168.6.150, length 46
02:04:06.672021 00:50:56:87:e3:6d > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.32 is-at 00:50:56:87:e3:6d, length 46
02:04:07.674726 2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.32 tell 192.168.6.150, length 46
02:04:07.674773 00:50:56:87:e3:6d > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.32 is-at 00:50:56:87:e3:6d, length 46
02:04:08.676733 2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.32 tell 192.168.6.150, length 46
02:04:08.676868 00:50:56:87:e3:6d > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.32 is-at 00:50:56:87:e3:6d, length 46
02:04:10.673678 2a:f0:5f:ae:c9:8b > ff:ff:ff:ff:ff:ff, ethertype ARP (0x0806), length 60: Request who-has 192.168.6.32 tell 192.168.6.150, length 46
02:04:10.674026 00:50:56:87:e3:6d > 2a:f0:5f:ae:c9:8b, ethertype ARP (0x0806), length 60: Reply 192.168.6.32 is-at 00:50:56:87:e3:6d, length 46

vmbr0上發現了2個request包和1個reply包
tap401000001i0上僅發現了request包，並且length不一致，未發現reply包。
NAS上一切正常
看到的現象就是VM在發送icmp以前須要先得到NAS的MAC地址，因此須要先經過ARP廣播獲取到NAS的MAC地址。

• P的vmbr0異常：接收到2個request，1個reply
• P的vmbr0異常：接收到reply未發送至tap401000001i0
  因爲以上2個緣由，致使VM沒法收到reply，因此icmp不能進行，ping失敗。

問題定位在數據鏈路層的ARP協議上。
在VM上執行arping -c 10 192.168.6.40
現象依據。再次證實以上2個現象。

# brctl show
bridge name bridge id       STP enabled interfaces
vmbr0       8000.ac1f6b342094   no      bond0
                            tap401000001i0
vmbr1       8000.74a4b500e768   no      bond1
# brctl showstp  vmbr0
vmbr0
 bridge id      8000.ac1f6b342094
 designated root    8000.ac1f6b342094
 root port         0            path cost          0
 max age          20.00         bridge max age        20.00
 hello time        2.00         bridge hello time      2.00
 forward delay         0.00         bridge forward delay       0.00
 ageing time         300.00
 hello timer           0.00         tcn timer          0.00
 topology change timer     0.00         gc timer           4.18
 flags


bond0 (1)
 port id        8001            state            forwarding
 designated root    8000.ac1f6b342094   path cost          4
 designated bridge  8000.ac1f6b342094   message age timer      0.00
 designated port    8001            forward delay timer    0.00
 designated cost       0            hold timer         0.00
 flags

tap401000001i0 (2)
 port id        8002            state            forwarding
 designated root    8000.ac1f6b342094   path cost        100
 designated bridge  8000.ac1f6b342094   message age timer      0.00
 designated port    8002            forward delay timer    0.00
 designated cost       0            hold timer         0.00
 flags

# brctl showmacs vmbr0
port no mac addr        is local?   ageing timer
  1 2a:f0:5f:ae:c9:8b   no         1.13
  1 00:50:56:87:e3:6d   no         0.00
# iptables -L
Chain INPUT (policy ACCEPT)
target     prot opt source               destination

Chain FORWARD (policy ACCEPT)
target     prot opt source               destination

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination
# ebtables -L
Bridge table: filter

Bridge chain: INPUT, entries: 0, policy: ACCEPT

Bridge chain: FORWARD, entries: 0, policy: ACCEPT

Bridge chain: OUTPUT, entries: 0, policy: ACCEPT
# ip rule show
0:  from all lookup local
32766:  from all lookup main
32767:  from all lookup default
# ip route show table all
default via 192.168.6.1 dev vmbr0 onlink
10.1.1.0/24 dev vmbr1 proto kernel scope link src 10.1.1.63
192.168.6.0/24 dev vmbr0 proto kernel scope link src 192.168.6.63
broadcast 10.1.1.0 dev vmbr1 table local proto kernel scope link src 10.1.1.63
local 10.1.1.63 dev vmbr1 table local proto kernel scope host src 10.1.1.63
broadcast 10.1.1.255 dev vmbr1 table local proto kernel scope link src 10.1.1.63
broadcast 127.0.0.0 dev lo table local proto kernel scope link src 127.0.0.1
local 127.0.0.0/8 dev lo table local proto kernel scope host src 127.0.0.1
local 127.0.0.1 dev lo table local proto kernel scope host src 127.0.0.1
broadcast 127.255.255.255 dev lo table local proto kernel scope link src 127.0.0.1
broadcast 192.168.6.0 dev vmbr0 table local proto kernel scope link src 192.168.6.63
local 192.168.6.63 dev vmbr0 table local proto kernel scope host src 192.168.6.63
broadcast 192.168.6.255 dev vmbr0 table local proto kernel scope link src 192.168.6.63
# bridge vlan show
port    vlan ids
bond0    1 PVID Egress Untagged

vmbr0    1 PVID Egress Untagged

bond1    1 PVID Egress Untagged

vmbr1    1 PVID Egress Untagged

tap401000001i0   1 PVID Egress Untagged

以上一切正常，未發現輸出問題，思路該怎麼展開？

google了一下午加一個晚上，各類不解決。
最後看到了linux關於bond的說明：
https://forum.huawei.com/enterprise/zh/thread-282727.html
再看看P的網卡，明明是mode0啊！
這時候再去交換機上一看！哈啊~！

interface GigabitEthernet0/0/5
 port link-type access
 port default vlan 6
#
interface GigabitEthernet0/0/6
 port link-type access
 port default vlan 6
#
interface GigabitEthernet0/0/7
 port link-type access
 port default vlan 6
#
interface GigabitEthernet0/0/8
 port link-type access
 port default vlan 6

感受好像是這個問題。
而後ifdown ens1f1，問題解決。
VM與全部外部網絡通信正常，無任何故障現象。
經過閱讀華爲網站的文章，肯定了balance-rr須要交換機作修改。
不想勞煩網管，因此本身改爲mode6。
一切歸於平靜。

問題只是臨時解決了，可是產生現象的原理還沒明白。

2009年本身就詳細看了windows teaming linux bond的原理及配置，苦於只是知道OS層面，不知道switch側應該怎麼對應，因此知識點一直有欠缺，2017年、2018年都曾經入坑，惋惜哪裏有2臺物理服務器和物理交換機實踐一下呀！
此次總算補齊了OS和switch的配置。

能用mode6就用mode6，再其次就是mode4。

如下給本身看的:

brctl show                          #查看bridge信息
brctl showstp  vmbr0                #查看vnbr0的stp信息，是否轉發等
brctl showmacs vmbr0 | grep c9:8b   #查看bridge的MAC地址
tcpdump -leni vmbr0 arp             #debug vmbr0網卡的arp包信息
tcpdump -leni vmbr0 icmp            #debug vmbr0網卡的icmp包信息
iptables -A FORWARD -i vmbr0 -o vmbr0 -j ACCEPT
iptables -L     #查看iptables
ebtables -L     #查看數據鏈路層的table信息
arping -c 10 192.168.1.1            #arp ping
ip rule show
ip route show table local
ip route show table all
bridge monitor          #monitor fdb update
bridge vlan show        #查看bridge的vlan信息

ping -c 1 -I veth1 192.168.3.1      #指定網卡進行ping測試

/proc/sys/net/bridge/bridge-nf-call-iptables    #幹嗎的
/proc/sys/net/ipv4/ip_forward                   #多網卡下，數據包轉發

sysctl -a | grep bridge
net.bridge.bridge-nf-call-arptables中的0或者1表明什麼意思？？

echo 1 > /proc/sys/net/ipv4/conf/eth0/arp_ignore        #什麼意思？？
echo 8 > /proc/sys/net/ipv4/conf/eth0/arp_announce      #什麼意思？？