[j-nsp] IPv6 MTU issue

[Gökhan Gümüş]

Dear Alex,

Thanks for the response.
Yes, you are right.
This service is configured on MX on logical router.
There is an aggregated interface is trunked to a switch as seen below,

abc> show configuration logical-systems tcr1 interfaces ae2.39
vlan-id 39;
family inet {
    address a.b.c.d/30;
}
family inet6 {
    address ..../124;

Do you think that just to increase INET6 MTU to 1504 bytes resolves this
issue?

Thanks,
Gokhan


On Fri, Oct 21, 2011 at 1:51 PM, Alex <alex.arseniev at gmail.com> wrote:

> My wild guess is that you are hitting a network of FreBSD/Olive/Linux IPv6
> routers with certain cards where when 802.1Q is enabled, MTU is
> automatically reduced by 4 bytes.
> You can simulate it with Olive by enabling "vlan-tagging" on Intel PRO/100
> card.
> My $0.02
> Thanks
> Alex
>
>
> ----- Original Message ----- From: "Gökhan Gümüs" <ggumus at gmail.com>
> To: <juniper-nsp at puck.nether.net>
> Sent: Friday, October 21, 2011 11:25 AM
> Subject: [j-nsp] IPv6 MTU issue
>
>
>  Dear all,
>>
>> I have an issue with IPv6 MTU.
>>
>> When i make a traceroute to one destination, the MTU size is reduced from
>> 1500 to 1496, somehow...?
>>
>> What makes it worse is that hop which drops the MTU does not send ICMP
>> "packet too long" message back..
>>
>> gg# scamper -I "trace -P udp -M abc
>> traceroute from xxx to abc
>> 1  s  0.192 ms [mtu: 1500]
>> 2  d  0.356 ms [mtu: 1500]
>> *3  f  1.422 ms [mtu: 1500]*
>> *  4  g  1.107 ms [*mtu: 1496]*
>>
>> 5  h  1.145 ms [*mtu: 1496]
>>
>>
>> On the link where MTU drops, output is as follows,
>>
>> Protocol inet, MTU: 1500
>>
>> Protocol inet6, MTU: 1500
>>
>>
>>
>> Software version is 10.0R4.7.
>>
>>
>>
>> Anybody has an experience on this behaviour before.
>>
>>
>>
>> Thanks and regards,
>>
>> Gokhan
>> ______________________________**_________________
>> juniper-nsp mailing list juniper-nsp at puck.nether.net
>> https://puck.nether.net/**mailman/listinfo/juniper-nsp<https://puck.nether.net/mailman/listinfo/juniper-nsp>
>>
>>
>