Home » General » Discussions » IPv6 auto-configuration issue with virtual Ethernet (Mixing IPv4 and IPv6 containers)
| Re: IPv6 auto-configuration issue with virtual Ethernet [message #41241 is a reply to message #41222] |
Thu, 16 December 2010 00:34   |
lars.bailey
Messages: 38
Registered: April 2010
|
Member |
|
|
Since my thread seems to be generating some interest,I decided to expanded it a little.
Actually,this is probably what I should have posted from the beginning.
To start,we are currently in the process of transitioning from IPv4 to IPv6.
A testbed network was setup that consists of a IPv4/IPv6 edge router(PC),and one OVZ Node server.(dual-stack)
The edge router,is configured to drop all outbound IPv6 packets,and accept IPv6 ICMP packets from the test network.
Two private "/64" IPv6 sub-nets,was used for the testbed.
One for internal IPv6 networking,and one for auto-configured IPv6.
I configured one Ethernet bridge on the Node server,which serves as the IPv4-IPv6 gateway,and proper routing added.
VZBR0 routing
192.168.200.0/24 scope link
192.168.254.0/24 proto kernel scope link src 192.168.254.4
fd60:1014:9458:4b60::/64 metric 1 mtu 1500 advmss 1440 hoplimit 4294967295
fd98:f0bd:b577:3c8b::/64 proto kernel metric 256 mtu 1500 advmss 1440 hoplimit 4294967295
fe80::/64 proto kernel metric 256 mtu 1500 advmss 1440 hoplimit 4294967295
RADVD was configured to advertise the selected "/64",on the Ethernet bridge.
Advertised prefix fd60:1014:9458:4b60::/64
In creating and configuring a OVZ test container,we used a in-house application,that builds a VE automatically,based on
the configuration input and the OS template selected.
All OS specific network configuration files,are removed during build,then a new one is created and proper,via the script.
(i.e "/etc/network,/etc/hosts,you get the picture)
So,when a VE container is configured for IPv4 only,it is built to support IPv4 only.
For IPv6,the same principle is applied.(static,dual-homed,auto-configure)
Five IPv4 containers was created,configured,and added to the bridge,and based on Fedora 13.
Five IPv6 auto-configured containers,was also created and added to the bridge,based on fedora 13.
In checking the network setup(s) for VE 4101,you get this;
/etc/sysconfig/network-scripts/ifcfg-eth0
DEVICE=eth0
TYPE=Ethernet
ONBOOT=yes
IPADDR=192.168.100.101
NETMASK=255.255.255.0
GATEWAY=192.168.100.101
/etc/sysconfig/network
HOSTNAME="moe.lan04.intranet"
GATEWAYDEV=eth0
This is the current network ouput for 4101;
[root@stooge network-scripts]# vzctl enter 4101
entered into VE 4101
[root@moe /]# ifconfig eth0
eth0 Link encap:Ethernet HWaddr xx:xx:xx:xx:xx:xx
inet addr:192.168.100.101 Bcast:192.168.100.255 Mask:255.255.255.0
inet6 addr: fd60:1014:9458:4b60:218:51ff:fe80:6450/64 Scope:Global
inet6 addr: fe80::218:xxxx:fe80:6450/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:5837 errors:0 dropped:0 overruns:0 frame:0
TX packets:11 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:560200 (547.0 KiB) TX bytes:560 (560.0 b)
VE4101 IPv4 routing
[root@moe /]# ip -4 ro show dev eth0
192.168.100.0/24 proto kernel scope link src 192.168.100.101
169.254.0.0/16 scope link
default via 192.168.100.101
VE4101 IPv6 routing
[root@moe /]# ip -6 ro show dev eth0
fd60:1014:9458:4b60::/64 proto kernel metric 256 expires 2147154sec mtu 1500 advmss 1440 hoplimit 4294967295
fe80::/64 proto kernel metric 256 mtu 1500 advmss 1440 hoplimit 4294967295
default via fe80::214:xxxx:fe5e:513f proto kernel metric 1024 expires 27sec mtu 1500 advmss 1440 hoplimit 64
[root@moe /]#
Here,VE4101 shows global IPv4 connectivity.
[root@moe /]# ping -c 3 www.openvz.org
PING www.openvz.org (64.131.90.7) 56(84) bytes of data.
64 bytes from openvz.org (64.131.90.7): icmp_seq=1 ttl=47 time=789 ms
64 bytes from openvz.org (64.131.90.7): icmp_seq=2 ttl=47 time=38.4 ms
64 bytes from openvz.org (64.131.90.7): icmp_seq=3 ttl=47 time=38.7 ms
--- www.openvz.org ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2042ms
rtt min/avg/max/mdev = 38.485/289.027/789.809/354.106 ms
[root@moe /]#
As you can tell from the IPv6 routing table,VE4101 has a default link-local route to the Ethernet bridge.
Yet,VE 4101 does not have any reference to any IPv6 specific configuration.
Hmmm.
The question that comes to my mind is whether or not,this same issue will arise on a physically network segment.
I guess,this will require additional testing on my part.
In conclusion,I really do not see at this time,this is an OVZ specific issue,and I really do not see a practical use of auto-configured IPv6 for containers.
The fix,was simply segregating IPv4 from IPv6,using an additional bridged interface.
|
|
|
|
Current Time: Sat Aug 31 10:16:53 GMT 2024
Total time taken to generate the page: 0.04585 seconds
|
OpenVZ Forum
Members
Search
Help
Register
Login
Home
