[c-nsp] Would an MTU mis-match cause one-way ICMP over EoMPLS VC?

[Manu Chao]

wrong mtu setting = normal problem, normal drop ;)

you must have the same mtu on a ptp link, if not fragmentation will fail

On Mon, Jun 29, 2009 at 6:17 AM, Jason Lixfeld <jason at lixfeld.ca> wrote:

> Diagram:
>
>  siteA CE
>    ||
> +---++---+
> | 7206PE |
> +---++---+
>   f2/0 (mtu 1500)
>    ||
>   f0/1 (mtu 1504)
> +---++---+
> | ME3400 |
> +---++---+
>   g0/1 (mtu 1504)
>    ||
>   g1/1 (mtu 9216)
> +---++---+
> |  7609  |
> +---++---+
>   g7/2 (mtu 9216)
>    ||
>   g0/0 (mtu 9216)
> +---++---+
> | 7301PE |
> +---++---+
>    ||
>  siteB CE
>
> I'm getting one-way ICMP over a VC that is terminated on the 7206PE;
> meaning ICMP echo requests sourced from siteA CE to siteB CE cannot be seen
> on the siteB CE.  However, ICMP echo requests sourced from the siteB CE can
> be seen on the siteA CE (but the echo reply packest are not seen by siteB
> CE).
>
> I understand that MTU issues would most certainly cause problems if the
> packet size was closer to the 1500 byte mark (1474 or there about,
> depending, maybe), but would this particular MTU mis-match even cause issues
> with such small ICMP packets?
>
> If MTU wouldn't cause this, then I'm back to square one with trying to
> figure out this one-way traffic thing I've got going on here.
>
> Thanks in advance..
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
>