KVM/Libvirt bridged/routed networking not working on newer guest kernels
I have a dedicated server running Debian 6, with Libvirt (0.9.11.3) and Qemu-KVM (qemu-kvm-1.0+dfsg-11, Debian).
I am having a problem getting bridged/routed networking to work in KVM guests with newer kernels (>2.6.38). NATted networking works fine though.
Older kernels work perfectly fine as well.
The host kernel is at version 3.2.0-2-amd64, the problem was also there on an older host kernel.
The contents of the host's /etc/network/interfaces (ip removed):
# Loopback device: auto lo iface lo inet loopback # bridge auto br0 iface br0 inet static address 176.9.xx.xx broadcast 176.9.xx.xx netmask 255.255.255.224 gateway 176.9.xx.xx pointopoint 176.9.xx.xx bridge_ports eth0 bridge_stp off bridge_maxwait 0 bridge_fd 0 up route add -host 176.9.xx.xx dev br0 # VM IP post-up mii-tool -F 100baseTx-FD br0 # default route to access subnet up route add -net 176.9.xx.xx netmask 255.255.255.224 gw 176.9.xx.xx br0
The output of ifconfig -a on the host:
br0 Link encap:Ethernet HWaddr 54:04:a6:8a:66:13
inet addr:176.9.xx.xx Bcast:176.9.xx.xx Mask:255.255.255.224
inet6 addr: fe80::5604:a6ff:fe8a:6613/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:20216729 errors:0 dropped:0 overruns:0 frame:0
TX packets:19962220 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:14144528601 (13.1 GiB) TX bytes:7990702656 (7.4 GiB)
eth0 Link encap:Ethernet HWaddr 54:04:a6:8a:66:13
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:26991788 errors:0 dropped:12066 overruns:0 frame:0
TX packets:19737261 errors:270082 dropped:0 overruns:0 carrier:270082
collisions:1686317 txqueuelen:1000
RX bytes:15459970915 (14.3 GiB) TX bytes:6661808415 (6.2 GiB)
Interrupt:17 Memory:fe500000-fe520000
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:6240133 errors:0 dropped:0 overruns:0 frame:0
TX packets:6240133 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:6081956230 (5.6 GiB) TX bytes:6081956230 (5.6 GiB)
virbr0 Link encap:Ethernet HWaddr 52:54:00:79:e4:5a
inet addr:192.168.100.1 Bcast:192.168.100.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:225016 errors:0 dropped:0 overruns:0 frame:0
TX packets:412958 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:16284276 (15.5 MiB) TX bytes:687827984 (655.9 MiB)
virbr0-nic Link encap:Ethernet HWaddr 52:54:00:79:e4:5a
BROADCAST MULTICAST MTU:1500 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:500
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
vnet0 Link encap:Ethernet HWaddr fe:54:00:93:4e:68
inet6 addr: fe80::fc54:ff:fe93:4e68/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:607670 errors:0 dropped:0 overruns:0 frame:0
TX packets:5932089 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:500
RX bytes:83574773 (79.7 MiB) TX bytes:1092482370 (1.0 GiB)
vnet1 Link encap:Ethernet HWaddr fe:54:00:ed:6a:43
inet6 addr: fe80::fc54:ff:feed:6a43/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:922132 errors:0 dropped:0 overruns:0 frame:0
TX packets:6342375 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:500
RX bytes:251091242 (239.4 MiB) TX bytes:1629079567 (1.5 GiB)
vnet2 Link encap:Ethernet HWaddr fe:54:00:0d:cb:3d
inet6 addr: fe80::fc54:ff:fe0d:cb3d/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:9461 errors:0 dropped:0 overruns:0 frame:0
TX packets:665189 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:500
RX bytes:4990275 (4.7 MiB) TX bytes:49229647 (46.9 MiB)
vnet3 Link encap:Ethernet HWaddr fe:54:cd:83:eb:aa
inet6 addr: fe80::fc54:cdff:fe83:ebaa/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:1649 errors:0 dropped:0 overruns:0 frame:0
TX packets:12177 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:500
RX bytes:77233 (75.4 KiB) TX bytes:2127934 (2.0 MiB)
The guest's /etc/network/interfaces, in this case running Ubuntu 12.04 (ip removed):
# This file describes the network interfaces available on your system
# and how to activate them. For more information, see interfaces(5).
# The loopback network interface
auto lo
iface lo inet loopback
auto eth0
iface eth0 inet static
address 176.9.xx.xx
netmask 255.255.255.248
gateway 176.9.xx.xx # Host IP
pointopoint 176.9.xx.xx # Host IP
dns-nameservers 8.8.8.8 8.8.4.4
The output of ifconfig -a on the guest:
eth0 Link encap:Ethernet HWaddr 52:54:cd:83:eb:aa
inet addr:176.9.xx.xx Bcast:0.0.0.0 Mask:255.255.255.255
inet6 addr: fe80::5054:cdff:fe83:ebaa/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:14190 errors:0 dropped:0 overruns:0 frame:0
TX packets:1768 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:2614642 (2.6 MB) TX bytes:82700 (82.7 KB)
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:954 errors:0 dropped:0 overruns:0 frame:0
TX packets:954 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:176679 (176.6 KB) TX bytes:176679 (176.6 KB)
Output of ping -c4 on the guest:
PING google.nl (173.194.35.151) 56(84) bytes of data. 64 bytes from muc03s01-in-f23.1e100.net (173.194.35.151): icmp_req=1 ttl=55 time=14.7 ms From static.174.82.xx.xx.clients.your-server.de (176.9.xx.xx): icmp_seq=2 Redirect Host(New nexthop: static.161.82.9.176.clients.your-server.de (176.9.82.161)) 64 bytes from muc03s01-in-f23.1e100.net (173.194.35.151): icmp_req=2 ttl=55 time=15.1 ms From static.198.170.9.176.clients.your-server.de (176.9.170.198) icmp_seq=3 Destination Host Unreachable From static.198.170.9.176.clients.your-server.de (176.9.170.198) icmp_seq=4 Destination Host Unreachable --- google.nl ping statistics --- 4 packets transmitted, 2 received, +2 errors, 50% packet loss, time 3002ms rtt min/avg/max/mdev = 14.797/14.983/15.170/0.223 ms, pipe 2
The static.174.82.xx.xx.clients.your-server.de (176.9.xx.xx) is the host's IP.
I have encountered this problem with every guest OS I've tried, that being Fedora, Ubuntu (server/desktop) and Debian with an upgraded kernel. I've also tried compiling the guest kernel myself, to no avail.
I have no problem with recompiling a kernel, though the host cannot afford any downtime.
Any ideas on this problem are very welcome.
EDIT: I can ping the host from inside the guest.
Solution 1:
Ok, I have found the solution myself. Apparently, the article on the Hetzner wiki wasn't complete. Redirects were still being sent, resulting in Hetzner (where we have our server) blocking the traffic. I am not sure how it had to do with what kernel version the guests were running, if someone can explain that I would very much like to hear it.
You can find out if you suffer from the same problem by executing the following command:
for file in `find /proc/ -iname send_redirects`; do echo $file; cat $file; done
which outputs something like:
/proc/sys/net/ipv4/conf/all/send_redirects 0 /proc/sys/net/ipv4/conf/default/send_redirects 1 /proc/sys/net/ipv4/conf/lo/send_redirects 1 /proc/sys/net/ipv4/conf/eth0/send_redirects 0 /proc/sys/net/ipv4/conf/br0/send_redirects 1 /proc/sys/net/ipv4/conf/vnet1/send_redirects 1 /proc/sys/net/ipv4/conf/vnet0/send_redirects 1 /proc/sys/net/ipv4/conf/virbr0/send_redirects 1 /proc/sys/net/ipv4/conf/virbr0-nic/send_redirects 1 /proc/sys/net/ipv4/conf/vnet2/send_redirects 1 /proc/sys/net/ipv4/conf/vnet3/send_redirects 1
I have to admit I do not know which ones should or should not be 0, but I have set them all to 0, and now it works fine, so I'll leave it at that.
for file in `find /proc/ -iname send_redirects`; do echo 0 > $file; done
Solution 2:
I had the same issue, thanks for sharing your solution. I think the setting is needed only on physical and bridge interface (eth0 and br0). You can make the change persistent by setting it into /etc/sysctl.d/
#/etc/sysctl.d/10-no-icmp-redirects.conf
# Hetzner says:
# Because of our network setup, the Host machine could send ICMP
# "redirect" messages to all guests, telling them to find the Hetzner
# gateway directly. That is impossible: Hetzner would throw away the
# traffic from the virtual interfaces because of their non registered
# MAC addresses (ie different from the main interface).
net.ipv4.conf.all.send_redirects=0
net.ipv4.conf.eth0.send_redirects=0
net.ipv4.conf.br0.send_redirects=0
net.ipv4.conf.default.send_redirects=0