kvm 客户机上的随机临时网络冻结

Ric*_*shi 4 networking linux bridge pause kvm-virtualization

我在 KVM 下运行两台来宾机器。主机运行的是 Arch Linux(内核版本 2.6.38.2),来宾都是 CentOS 5.5(内核版本 2.6.18)。我将网络设置为桥梁,或多或少如下(当然,我不是手工完成的——它分散在系统启动脚本中)

# /usr/sbin/brctl addbr br0
# /usr/sbin/brctl addif br0 eth0
# /sbin/ifconfig br0 192.168.3.30 netmask 255.255.255.0 broadcast 192.168.3.255 up
# qemu-kvm -hda <disk-image1> -m 2048 -vga std -vnc :3 -net nic,vlan=0 -net tap,vlan=0,ifname=tap0,script=/etc/qemu-ifup
# qemu-kvm -hda <disk-image2> -m 2048 -vga std -vnc :4 -net nic,vlan=0 -net tap,vlan=0,ifname=tap1,script=/etc/qemu-ifup
# cat /etc/qemu-ifup
#!/bin/sh
sudo /sbin/ifconfig $1 0.0.0.0 promisc up
sudo /usr/sbin/brctl addif br0 $1
Run Code Online (Sandbox Code Playgroud)

该配置主要工作。奇怪的是我经常出现网络“暂停”。我与虚拟机的大部分交互都是通过 ssh 或 Web 浏览器进行的。通常,特别是在空闲时间之后,我会在会话中输入一些内容或访问网页,但什么也不会发生。最终,10-15 秒后,正常活动将恢复正常。从死时间开始的 ssh 击键会匆匆忙忙地完成。Web 请求似乎丢失了,但是快速的“重新加载页面”立即得到了结果。整个虚拟机一直在运行,正如我通过观察控制台的 VNC 会话所看到的。两个客人就这样独立了。有时两者会同时冻结,但有时一个工作正常,而另一个冻结。

关于解决方案、原因或进一步诊断的任何想法?


(补充信息,7/27)

这是主机上“ifconfig -a”的输出:

hope 5$ sudo ifconfig -a
br0       Link encap:Ethernet  HWaddr 4A:F0:CA:32:F5:88  
          inet addr:192.168.3.30  Bcast:192.168.3.255  Mask:255.255.255.0
          inet6 addr: fe80::7271:bcff:fea2:f4c4/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:140947463 errors:0 dropped:0 overruns:0 frame:0
          TX packets:66981660 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:192776331275 (183845.8 Mb)  TX bytes:9469675717 (9030.9 Mb)

br0:1     Link encap:Ethernet  HWaddr 4A:F0:CA:32:F5:88  
          inet addr:192.168.3.31  Bcast:192.168.3.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

br0:2     Link encap:Ethernet  HWaddr 4A:F0:CA:32:F5:88  
          inet addr:192.168.3.32  Bcast:192.168.3.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

eth0      Link encap:Ethernet  HWaddr 70:71:BC:A2:F4:C4  
          inet6 addr: fe80::7271:bcff:fea2:f4c4/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:141848164 errors:0 dropped:0 overruns:0 frame:0
          TX packets:68852413 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:195831275369 (186759.2 Mb)  TX bytes:9559920997 (9117.0 Mb)
          Interrupt:20 Memory:d0300000-d0320000 

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:27197534 errors:0 dropped:0 overruns:0 frame:0
          TX packets:27197534 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:127242745992 (121348.1 Mb)  TX bytes:127242745992 (121348.1 Mb)

sit0      Link encap:UNSPEC  HWaddr 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00  
          NOARP  MTU:1480  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)

tap0      Link encap:Ethernet  HWaddr 4A:F0:CA:32:F5:88  
          inet6 addr: fe80::48f0:caff:fe32:f588/64 Scope:Link
          UP BROADCAST RUNNING PROMISC MULTICAST  MTU:1500  Metric:1
          RX packets:397512 errors:0 dropped:0 overruns:0 frame:0
          TX packets:668318 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:500 
          RX bytes:38045826 (36.2 Mb)  TX bytes:750358170 (715.5 Mb)

tap1      Link encap:Ethernet  HWaddr 56:92:5E:DE:93:67  
          inet6 addr: fe80::5492:5eff:fede:9367/64 Scope:Link
          UP BROADCAST RUNNING PROMISC MULTICAST  MTU:1500  Metric:1
          RX packets:25635 errors:0 dropped:0 overruns:0 frame:0
          TX packets:88846 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:500 
          RX bytes:1999805 (1.9 Mb)  TX bytes:17071986 (16.2 Mb)
Run Code Online (Sandbox Code Playgroud)

主机上“brctl show”的输出:

hope 6$ sudo brctl show
bridge name bridge id               STP enabled     interfaces
br0         8000.4af0ca32f588       no              eth0
                                                    tap0
                                                    tap1
Run Code Online (Sandbox Code Playgroud)

'ifconfig -a' 在一个来宾上的输出:

guest1# ifconfig -a
eth0      Link encap:Ethernet  HWaddr 52:54:00:12:34:56  
          inet addr:192.168.3.35  Bcast:192.168.3.255  Mask:255.255.255.0
          inet6 addr: fe80::5054:ff:fe12:3456/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:323061 errors:0 dropped:0 overruns:0 frame:0
          TX packets:210150 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:298541685 (284.7 MiB)  TX bytes:25958962 (24.7 MiB)
          Interrupt:11 Base address:0x4000 

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:2305 errors:0 dropped:0 overruns:0 frame:0
          TX packets:2305 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:481780 (470.4 KiB)  TX bytes:481780 (470.4 KiB)

sit0      Link encap:IPv6-in-IPv4  
          NOARP  MTU:1480  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
Run Code Online (Sandbox Code Playgroud)

和另一位客人同样的事情:

guest2# ifconfig -a
eth0      Link encap:Ethernet  HWaddr 52:54:00:12:34:56  
          inet addr:192.168.3.36  Bcast:192.168.3.255  Mask:255.255.255.0
          inet6 addr: fe80::5054:ff:fe12:3456/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:62641 errors:0 dropped:0 overruns:0 frame:0
          TX packets:25718 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:10789140 (10.2 MiB)  TX bytes:2012261 (1.9 MiB)
          Interrupt:11 Base address:0x4000 

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:182 errors:0 dropped:0 overruns:0 frame:0
          TX packets:182 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:25798 (25.1 KiB)  TX bytes:25798 (25.1 KiB)

sit0      Link encap:IPv6-in-IPv4  
          NOARP  MTU:1480  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
Run Code Online (Sandbox Code Playgroud)

Ric*_*shi 5

我终于找到了这个问题的正确答案。

简短回答:MAC 地址冲突。

长答案:对于每个虚拟机,qemu 都会创建一个虚拟网卡。它为这个虚拟 NIC 分配一个 MAC 地址。默认情况下,它为每个 NIC 分配一个固定地址 (52:54:00:12:34:56)。如果您在同一主机上启动多个来宾,他们都会获得相同的固定地址。您甚至可以在我的原始问题的 ifconfig 输出中看到这一点。自然地,如果您将来宾桥接在一起,这是一个问题,因为这与在同一网络上拥有两台具有相同 MAC 地址的物理机相同。ARP 尽其所能并在它们之间来回切换,但超时时间很长,给了我我看到的结果。

因此,解决方案是为每个来宾添加“macaddr=...”标志到 -net 参数,这样每个来宾都有唯一的 MAC 地址。例子:

# qemu-kvm -hda <disk-image1> -m 2048 -vga std -vnc :3 -net nic,vlan=0,macaddr=52:54:00:00:00:03 -net tap,vlan=0,ifname=tap0,script=/etc/qemu-ifup
# qemu-kvm -hda <disk-image2> -m 2048 -vga std -vnc :4 -net nic,vlan=0,macaddr=52:54:00:00:00:04 -net tap,vlan=0,ifname=tap1,script=/etc/qemu-ifup
Run Code Online (Sandbox Code Playgroud)

在我看来,qemu 的默认行为应该更聪明,但我认为每个来宾每次启动时都获得相同的 MAC 地址是相当重要的,无论它们启动的顺序如何。也许作者想不到一种方法来确保这一点。

无论如何,如果您在同一网络(虚拟或其他)上运行多个来宾,请确保明确且唯一地设置 MAC 地址。否则,你会遇到同样的问题。