Question

我有RHEL6.4操作系统的主机。在主机下我有一个运行RHEL6.4的KVM VM。

主机连接到带有vlan标记1的标记交换机接口。

我已按如下方式配置主机网络接口：

的ifcfg-eth0的

＃

DEVICE=eth0
TYPE=Ethernet
ONBOOT=yes
USERCTL=no
BOOTPROTO=none
MASTER=bond0
SLAVE=yes

＃

的ifcfg-eth1的

＃

DEVICE=eth1
TYPE=Ethernet
ONBOOT=yes
USERCTL=no
BOOTPROTO=none
MASTER=bond0
SLAVE=yes

＃

的ifcfg-bond0

＃

DEVICE=bond0
TYPE=Bond
BONDING_MASTER=yes
BOOTPROTO=static
NAME=bond0
ONBOOT=yes
USERCTL=no
BONDING_OPTS="mode=active-backup primary=eth0 miimon=100"

＃

和主机vswitch如下：

＃

ovs-vsctl add-br ovsbr0
ovs-vsctl add-port ovsbr0 bond0
ovs-vsctl add-port ovsbr0 vlan1 tag=1 -- set interface content type=internal
ifconfig vlan1 192.168.1.2/28 up

＃

我能够ping外部vlan1 GW（192.168.1.1）。

我遇到的问题是我无法从VM ping外部IPv4网关。当我从VM ping时，我得到第一个数据包的响应，但其余的似乎丢失了VLAN标记并进入了错误的接口。如果我tcpdump bond0.1我看到第一个ICMP请求和回复，但其余的ICMP回复似乎出现在bond0接口（而不是bond0.1）

＃

[root@vm ~]# ping 192.168.1.1
PING 192.168.1.1 (192.168.1.1) 56(84) bytes of data.
64 bytes from 192.168.1.1: icmp_seq=1 ttl=255 time=2.64 ms
^C
--- 192.168.1.1 ping statistics ---
5 packets transmitted, 1 received, 80% packet loss, time 4786ms
rtt min/avg/max/mdev = 2.642/2.642/2.642/0.000 ms
[root@vm ~]#

＃

[root@vm ~]#  tcpdump -i bond0.1 host 192.168.1.1
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on bond0.1, link-type EN10MB (Ethernet), capture size 65535 bytes
20:49:28.053502 IP 192.168.1.3 > 192.168.1.1: ICMP echo request, id 55664, seq 1, length 64
20:49:28.054158 IP 192.168.1.1 > 192.168.1.3: ICMP echo reply, id 55664, seq 1, length 64
20:49:29.055189 IP 192.168.1.3 > 192.168.1.1: ICMP echo request, id 55664, seq 2, length 64
20:49:30.054971 IP 192.168.1.3 > 192.168.1.1: ICMP echo request, id 55664, seq 3, length 64
20:49:31.054935 IP 192.168.1.3 > 192.168.1.1: ICMP echo request, id 55664, seq 4, length 64
20:49:32.054918 IP 192.168.1.3 > 192.168.1.1: ICMP echo request, id 55664, seq 5, length 64

＃

[root@vm]# tcpdump -i bond0 host 192.168.1.1
tcpdump: WARNING: bond0: no IPv4 address assigned
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on bond0, link-type EN10MB (Ethernet), capture size 65535 bytes
20:49:28.053509 IP 192.168.1.3 > 192.168.1.1: ICMP echo request, id 55664, seq 1, length 64
20:49:29.055199 IP 192.168.1.3 > 192.168.1.1: ICMP echo request, id 55664, seq 2, length 64
20:49:29.055653 IP 192.168.1.1 > 192.168.1.3: ICMP echo reply, id 55664, seq 2, length 64
20:49:30.054983 IP 192.168.1.3 > 192.168.1.1: ICMP echo request, id 55664, seq 3, length 64
20:49:30.055522 IP 192.168.1.1 > 192.168.1.3: ICMP echo reply, id 55664, seq 3, length 64
20:49:31.054968 IP 192.168.1.3 > 192.168.1.1: ICMP echo request, id 55664, seq 4, length 64
20:49:31.056497 IP 192.168.1.1 > 192.168.1.3: ICMP echo reply, id 55664, seq 4, length 64
20:49:32.054927 IP 192.168.1.3 > 192.168.1.1: ICMP echo request, id 55664, seq 5, length 64
20:49:32.055402 IP 192.168.1.1 > 192.168.1.3: ICMP echo reply, id 55664, seq 5, length 64

VM网络接口配置如下：

的ifcfg-eth0的

＃

DEVICE=eth0
TYPE=Ethernet
ONBOOT=yes
USERCTL=no
BOOTPROTO=none
MASTER=bond0
SLAVE=yes

的ifcfg-eth1的

＃

DEVICE=eth1
TYPE=Ethernet
ONBOOT=yes
USERCTL=no
BOOTPROTO=none
MASTER=bond0
SLAVE=yes

的ifcfg-bond0

＃

DEVICE=bond0
TYPE=Bond
BONDING_MASTER=yes
BOOTPROTO=static
NAME=bond1
ONBOOT=yes
USERCTL=no
BONDING_OPTS="mode=active-backup primary=eth0 miimon=100"

ifcfg-bond0.1

＃

VLAN=yes
TYPE=Vlan
DEVICE=bond0.1
NAME=bond0.1
BOOTPROTO=static
ONBOOT=yes
IPADDR=192.168.1.3
NETMASK=255.255.255.240

知道为什么会出现这个问题吗？

Answer 1

默认情况下，Vlan 1未标记。因此，当您使用它作为标记（端口在干线模式）时，您应该在交换机上选择另一个未标记的（默认vlan）。在错误的vlan配置上，您可能会看到上述行为：您正在发送流量，但您无法收回流量。所以创建另一个vlan并再试一次。

同时使用this article

检查sysctl和其他设置

Openvswitch / KVM Guest无法ping外部GW

1 个答案: