[j-nsp] Logical Tunnels & IPv6

Mark Kamichoff prox at prolixium.com
Mon May 31 11:07:32 EDT 2010


Hi Chuck - 

On Sun, May 30, 2010 at 09:01:10PM -0400, Chuck Anderson wrote:
> Yes, and I believe the reason why this is the case is because 
> logical-tunnels use the same MAC address on each end.  Since IPv6 uses 
> the MAC address to generate the link-local address by default, that 
> may be why they prevent you from configuring inet6 on lt.

Actually, with my setup, it looks like the MACs are unique.  It
certainly reuses MAC addresses, but they are unique per "peering" of
logical interfaces.  For example:

{master}
lab at mx240-lab01-re0> show arp                
MAC Address       Address         Name                      Interface     Flags
[...snip]
00:22:83:32:cd:35 10.0.4.1        10.0.4.1                  lt-2/0/10.6   none
00:22:83:32:cd:34 10.0.4.2        10.0.4.2                  lt-2/0/10.7   none
00:22:83:32:cd:34 10.0.4.5        10.0.4.5                  lt-2/0/10.1   none
00:22:83:32:cd:35 10.0.4.6        10.0.4.6                  lt-2/0/10.0   none
[...snip]

lt-2/0/10.0 and lt-2/0/10.1 are connected, and so are lt-2/0/10.6 and
lt-2/0/10.7.  /30s.

For another data point, I've got an SRX 210 (10.1R1.8) doing something
similar, it just uses different addresses:

prox at orb> show arp no-resolve 
MAC Address       Address         Interface     Flags
00:26:88:e9:54:80 10.3.7.160      lt-0/0/0.1     none
00:26:88:e9:54:81 10.3.7.161      lt-0/0/0.0     none
00:26:88:e9:54:81 10.3.7.162      lt-0/0/0.2     none
00:26:88:e9:54:80 10.3.7.163      lt-0/0/0.3     none
[...snip]

/31s.

At least for this (simple) setup, it shouldn't be a reason why IPv6
can't be used.

- Mark

-- 
Mark Kamichoff
prox at prolixium.com
http://www.prolixium.com/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <https://puck.nether.net/pipermail/juniper-nsp/attachments/20100531/03a64f53/attachment-0001.bin>


More information about the juniper-nsp mailing list