[j-nsp] L2VPN MTU Issue

[Eric Van Tol]

> -----Original Message-----
> From: juniper-nsp-bounces at puck.nether.net [mailto:juniper-nsp-
> bounces at puck.nether.net] On Behalf Of Eric Van Tol
> Sent: Wednesday, September 29, 2010 2:57 PM
> To: juniper-nsp
> Subject: Re: [j-nsp] L2VPN MTU Issue
>
> Well, primary credit goes to JTAC, but doing some packet captures, I found
> that the packets entering the customer interface were not leaving the router.
> JTAC suggested 1520, but that only allowed 1498-bytes through.  I bumped it
> up to 1524, which still did not work, so then I bumped it up to 1528.  I am
> still not too keen on the math, but JTAC hasn't broken it down for me yet.
> If anyone out there knows, please point me to a reference.  I googled like
> mad yesterday, but either overlooked the obvious, or wasn't searching for the
> right info.
> 
> -evt

I was reading a partial thread yesterday in the archives and for whatever reason, could not find the reference thread and today, I came upon this:

http://puck.nether.net/pipermail/juniper-nsp/2006-September/007000.html

"In a typical l2 vpn there would be two labels added; vrf and outer transport. The core interfaces will need an mtu of ethernet (1500) + vlan (4) + martini encap (0 | 4) + labels
(8) + what ever link encap is needed based on core interface type. If the core interfaces are also ethernet then this is another 14 for total of 1526-1530 (the latter is with optional martini control word)"

I came up with 1504 + 8 + 14 = 1526 in this calculation.  I am 2 bytes over the necessary, but to be honest, I will probably bump it up to 1530 or higher, so that 802.1q encapsulation can be supported.  For anyone starting out in MPLS VPNs in the Juniper world, this thread is a good one to read, as Harry points out at least one (perhaps previously, now) undocumented function of the MPLS MTU later on in one of his responses.

Thanks, Harry!

-evt