<html><head><meta http-equiv="Content-Type" content="text/html charset=us-ascii"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class="">Here, the link local is on the tunnel, right? Unexpected things can happen. Can we be sure both ends have same MTU? What is the effect if they are different?<div class=""><br class=""></div><div class="">If you mean we should circumvent fragmentation, I fully agree.</div><div class=""><br class=""></div><div class="">Happy Christmas,</div><div class="">Teco<br class=""><div class=""><br class=""></div><div class=""><br class=""></div><div class=""><br class=""><div><blockquote type="cite" class=""><div class="">Op 24 dec. 2014, om 23:32 heeft Henning Rogge <<a href="mailto:hrogge@gmail.com" class="">hrogge@gmail.com</a>> het volgende geschreven:</div><br class="Apple-interchange-newline"><div class=""><p dir="ltr" class="">PMTUD for link local multicast?</p><p dir="ltr" class="">Henning</p>
<div class="gmail_quote">Am 24.12.2014 15:41 schrieb "Teco Boot" <<a href="mailto:teco@inf-net.nl" class="">teco@inf-net.nl</a>>:<br type="attribution" class=""><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">DF=1 and no PMTUD makes no sense.<br class="">
Having a function that circumvent fragments is great. But set DF and discard PTB messages is wrong.<br class="">
<br class="">
If both sides of a link have different MTU, there could still be a problem.<br class="">
<br class="">
Teco<br class="">
<br class="">
> Op 24 dec. 2014, om 12:48 heeft Henning Rogge <<a href="mailto:hrogge@gmail.com" class="">hrogge@gmail.com</a>> het volgende geschreven:<br class="">
><br class="">
> On Wed, Dec 24, 2014 at 10:46 AM, Jernej Kos <<a href="mailto:jernej@kos.mx" class="">jernej@kos.mx</a>> wrote:<br class="">
>> Hello!<br class="">
>><br class="">
>> The absence of HELLOs led me to investigate the MTUs more closely and it<br class="">
>> now seems that there is actually some issue with MTU discovery on the<br class="">
>> broker side -- nodes configure the correct MTU towards the broker, but<br class="">
>> the broker sometimes fails to configure correct MTU towards the nodes.<br class="">
>><br class="">
>> So this may be the cause of the problem and not OLSR as HELLOs from the<br class="">
>> server might then be dropped as being too large when received at the<br class="">
>> node. Will report back when I determine what is causing this.<br class="">
><br class="">
> Yes, I could see this becoming a problem.<br class="">
><br class="">
> olsrd uses the advertised MTU to make the packets as small as<br class="">
> necessary and mark them as "do not fragment"... if the MTU is wrong,<br class="">
> the packets might be dropped.<br class="">
><br class="">
> Henning Rogge<br class="">
><br class="">
> --<br class="">
> Olsr-users mailing list<br class="">
> <a href="mailto:Olsr-users@lists.olsr.org" class="">Olsr-users@lists.olsr.org</a><br class="">
> <a href="https://lists.olsr.org/mailman/listinfo/olsr-users" target="_blank" class="">https://lists.olsr.org/mailman/listinfo/olsr-users</a><br class="">
<br class="">
</blockquote></div>
</div></blockquote></div><br class=""></div></div></body></html>