Usually lower MTUs are not a problem for properly configured networks, as even lower MTU systems should begin fragmenting traffic. The issue comes where there is a mismatch, and suddenly, larger than expected packets are dropped. <br><br><br>So I think the tor protocol does not need to support identifying relay MTUs, but I'm also unsure if it tests the largest cell size against relays either. Testing a very large cell size should identify if a relay is properly configured. <br><br><br><br>-------- Original Message --------<br>On Feb 22, 2024, 5:47 AM, s7r - s7r at sky-ip.org < s7r@sky-ip.org> wrote:<blockquote class="protonmail_quote"><br>pasture_clubbed242--- via tor-relays wrote:
> Greetings,
>
> I believe there is a larger sized guard relay that has been having MTU
> issues for about a week. All connections with packets above a certain
> size are dropped. This results in partially loaded or broken webpages,
> broken file downloads, etc. Do Tor directory authorities test MTU
> (implicitly by speed test?) when testing relays?
>
> Wondering if anyone else noticed this or if it would be handled
> automatically by dir authorities.
>
> Thanks all
>

This is indeed very interesting. I never experienced this problem but
now that you mention it I will setup a test environment with some non
standard MTU values. I doubt the directory authorities test also the
MTU, but it's an interesting question, let's hope someone hosting a
bandwidth authority will reply to this.

Also, I'm not sure and I'm very curios what the bandwidth authorities
should do about this? What if a relay has super good speed but very low
MTU? Should it be excluded and marked as not running? Because it will be
very hard for Tor to also include MTU in the router descriptors and be
aware about it.

_______________________________________________
tor-relays mailing list
tor-relays@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
</div>