Hi Mitch,
This is my /etc/network/interfaces file from my lab server running Ubuntu
12.04 Server. I use two NICs, the built-in Intel gig NIC (eth0) for the
breakout (since it will do MTU larger than 1500) and a USB NIC (eth1) for
my access to the box from my internal LAN. I found the same as you, that
you can't leave an interface without an IP and actually have it come up. I
don't remember if I tried IP's in 127/8, but I'm OK burning a few IPs out
of 1.1.1.0/24 for my lab box. Note the 'ethtool' bits as annotated below.
Apparently newer Linux kernels will try to use VLAN tag offloading (similar
to TCP offload, etc) that will strip all incoming VLAN tags (both layers if
it's double-tagged) and it was causing major connectivity problems. Frames
were getting out from the virtual routers but not back in. Someone in my
Twitter network was able to point me to this solution and it's worked
perfectly. Was not necessary on the same box in Ubuntu 11.04.
This setup works great for me. In my Dynagen .net files (I run text-mode
dynagen, no X and thus no GNS) I just tie the virtual router interfaces to
a "vlan1XX" interface using the NIO_linux_eth option, as such:
[[ROUTER R5]]
model = 3745
console = 2005
f0/0 = NIO_linux_eth:vlan109
f0/1 = NIO_linux_eth:vlan110
=============== /etc/network/interfaces ==================
# The primary network interface
auto eth1
iface eth1 inet static
address 10.4.111.250
netmask 255.255.255.0
gateway 10.4.111.1
dns-nameservers 10.4.11.11
dns-search mccouch.private
iface eth1 inet6 static
address 2001:0470:XXXX:XXXX:0000:0000:0000:0250
netmask 64
# This is the base interface for the VLAN trunk
auto eth0
iface eth0 inet static
address 1.1.1.255
netmask 255.255.255.255
# Need to bump up MTU to ensure 1500-byte QinQ frames make it OK.
mtu 1998
# This is necessary on recent Linux kernels as VLAN tag offloading clobbers
the QinQ behavior we need from Dynamips.
# I spent 3 days getting this solved after rebuilding my lab box to get it
from Ubuntu 11.04 to 12.04.
post-up /sbin/ethtool --offload eth0 rxvlan off && /sbin/ethtool --offload
eth0 txvlan off
auto vlan101
auto vlan102
auto vlan103
auto vlan104
auto vlan105
auto vlan106
auto vlan107
auto vlan108
auto vlan109
auto vlan110
auto vlan111
auto vlan112
auto vlan113
auto vlan114
auto vlan115
auto vlan116
auto vlan117
auto vlan118
auto vlan119
auto vlan120
auto vlan121
auto vlan122
auto vlan123
auto vlan124
# VLAN 101
iface vlan101 inet static
address 1.1.1.1
netmask 255.255.255.255
mtu 1500
vlan-raw-device eth0
# VLAN 102
iface vlan102 inet static
address 1.1.1.2
netmask 255.255.255.255
mtu 1500
vlan-raw-device eth0
# VLAN 103
iface vlan103 inet static
address 1.1.1.3
netmask 255.255.255.255
mtu 1500
vlan-raw-device eth0
# VLAN 104
iface vlan104 inet static
address 1.1.1.4
netmask 255.255.255.255
mtu 1500
vlan-raw-device eth0
...And on and on up through VLAN 124.
On Sat, Sep 29, 2012 at 9:07 PM, G. Mitchell Peterson <
[email protected]> wrote:
> I'm working on getting my lab up and running using the breakout switch
> method on an Ubuntu Server 12.04 x64 server. I can't seem to get the trunk
> on the server working though. Can someone post a snip-it of their
> interfaces file? I was hoping that it behaved like a dot1q trunk in Cisco
> world, but that doesn't seem to be the case. There's no trunk allowed type
> command from what I can gather. Do I just make a sub for each VLAN going
> over the trunk like below? I really, really dislike that I have to assign
> an IP when using the static command so I'm hoping one of you network
> wizards knows a trick to this. I tried only using auto ethX and setting
> the MTU but it error'd out. So, apparently either you assign an IP or you
> assign DHCP but you have to do one or the other. Anyone see an issue with
> assigning a /32 in the 127.0.0.0/8 network to these since I have to give
> it
> something?
>
> #Trunk to breakout
> auto ethX
> iface ethX inet static
> adddress 127.0.0.100
> netmask 255.255.255.255
> mtu 1536
>
> #VLAN 101 on ethX
> auto ethX.101
> iface ethX.101 inet static
> adddress 127.0.0.101
> netmask 255.255.255.255
>
> #VLAN 102 on ethX
> auto ethX.102
> iface ethX.102 inet static
> adddress 127.0.0.102
> netmask 255.255.255.255
>
> Thanks,
> Mitch
> _______________________________________________
> For more information regarding industry leading CCIE Lab training, please
> visit www.ipexpert.com
>
> Are you a CCNP or CCIE and looking for a job? Check out
> www.PlatinumPlacement.com
>
> http://onlinestudylist.com/mailman/listinfo/ccie_rs
>
_______________________________________________
For more information regarding industry leading CCIE Lab training, please visit
www.ipexpert.com
Are you a CCNP or CCIE and looking for a job? Check out
www.PlatinumPlacement.com
http://onlinestudylist.com/mailman/listinfo/ccie_rs
