Well this is odd... IPv6 ssh traffic just broke for me. ping still works, suggesting something somewhere fucked up with MTUs all of a sudden I still need to dig into this. Just wondered if anyone else was seeing this with ARP? Or a home HE.net tunnel for that matter? i've seen ipv6 mtu discovery not work have you tried mss clamping? mss is clamped as I recall What's weird is that "everything was working", meaning my SSH sessions were ticking along just fine. Then all of a sudden they hung... but not at the same time. (one hung at 12:09, the other at 12:10 according to the clock in tmux) so *something* *somehow* broke. they hung with large amounts of text? or typing? hmm my ipv6 tunnel seems to be working normally to arp my tunnel is tserv3.fmt2 mercutio: Neither. They were both just sitting idle at prompts. Noticed they'd hung when I tried to type, then looked at the clock in my tmux bar on each session and it had frozen. that doesn't sound mtu related unless something spat lots of text at it it sounds like what happens when something loses state for some reason and is enforcing state. More digging later - it's an HE issue and I can't hit anything IPv6 from my home tunnel. http://sprunge.us/TPPC Reverse mtr from vps3 back towards my home connection showing that it gets lost somewhere within HE http://sprunge.us/JgIR So it's an HE issue -- Sorry for the noise. (The pings as it turns out were hitting the wrong addresses earlier, oops) you're in fmt2 too? No? sea1 *And no, https://tunnelbroker.net/status.php says no issues) ( hmm i wonder what it is/was According to mtr, looks like all is well now. The missing node in both directions was 100ge14-1.core1.pdx1.he.net so I guess they had a brief routing issue. (or router issue, technically) Side-by-side comparison of mtrs in both directions, before and after http://sprunge.us/UcjD ahh brycec: do you use netflix over your he.net tunnel? LOL nope, no thanks to Netflix. nathani: Not since Netflix classified HE as a proxy service.