hm think ipv6 just hit the fan again (or still?) Studying my Smokeping, it looks to have returned for a 3rd night (or day, or whatever your timezone declares) The past 6 days as measured from a TWC cable connection + HE IPv6 tunnel to my ARP VPS, you see that the heavy packet loss is unique to the past 3 days and starts up around 5pm ARP Daylight Time. Traceroute show clean through HE up until the handoff to ntt.net; my VPS is the hop immediately after the handoff to ntt.net. https://smokeping.cobryce.com/?displaymode=n;start=2015-08-29%2008:39;end=now;target=ARP.ARPWebsite.ARPWebsite6~cobrycesmokepingslave with high loss? Yep So I'd have to conclude that someone is saturating the living hell out of that connection eek that loosk severe are they at the same time of night? Yes (as I already said) but i mean is the start time pretty consistent Yes "unique to the past 3 days and starts up around 5pm ARP Daylight Time." oh right you did say sorry :) And I'm seeing the loss from multiple ISPs (but all happen to be tunneled HE IPv6 connections; I don't have a native IPv6 connection to monitor from, besides ARP itself.) (and because they're all HE tunnels, they all cross that same handoff it seems) yeah the site that i noticed loss from seems to not be too bad atm gtt -> ntt seems fine now still seeing high pings from he.net on any2ix www.he.net is showing 30% loss and bad loss to that tunenl server hmm bryce: is it any better now? dropping right now yeah hmm didn't seem to work flowing again http://sprunge.us/hhMb mtr I wonder... I see from the rdns names that HE's hop is 10GB, and NTT's is 1GB. Latency spikes on both that last HE hop as well as the NTT hop, which suggests (not conclusively) to me that that the link is oversaturated, and HE can handle it simply because they've got more capacity. but that's just a hypothesis. ah well, i've got other things to deal with, things I can actually affect. Later all ok i'm seeing less loss now btw but slightly higher ping to the feed dead name