<div dir="ltr"><div>My understanding is well behaved hosts will send DF traffic only for PMTUD. If you're sending regular traffic with DF, you're pretty much asking for it.<br><br></div>Paul Wilkins<br></div><div class="gmail_extra"><br><div class="gmail_quote">On 3 June 2015 at 19:23, Jeremy Visser <span dir="ltr"><<a href="mailto:jeremy@sunriseroad.net" target="_blank">jeremy@sunriseroad.net</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">On 03/06/15 17:03, Paul Wilkins wrote:<br>
> I wonder if Amazon are not RFC 791 compliant, because 'This usually<br>
> signifies an MTU misconfiguration on the remote end, we cannot help<br>
> with this' suggests a problem with packets dropping when oversize,<br>
> rather than fragmenting?<br>
<br>
</span>Most hosts speaking TCP set the DF (Don’t Fragment) bit, thereby encouraging MTU boundaries to send ICMP fragmentation reports. (Feature, not a bug.)<br>
<br>
The alternative is to speak TCP without the DF bit set, thereby causing fragmentation at an MTU boundary, leading to worse efficiency.<br>
<br>
Both situations suck, because both ICMP and fragmented packets are subject to senseless amounts of filtering by uninformed admins.<br>
<br>
In IPv6, fragmentation is end-to-end, making it even more crucially important for intermediate hops to not block vital ICMP messages. (Dear Internet: ICMP is not optional.)<br>
<div class="HOEnZb"><div class="h5">_______________________________________________<br>
AusNOG mailing list<br>
<a href="mailto:AusNOG@lists.ausnog.net">AusNOG@lists.ausnog.net</a><br>
<a href="http://lists.ausnog.net/mailman/listinfo/ausnog" target="_blank">http://lists.ausnog.net/mailman/listinfo/ausnog</a><br>
</div></div></blockquote></div><br></div>