On my VDS, both 6to4 and Teredo seem to work poorly (packet loss is about 40%).
just-ping.com shows packet loss 10%-80% both for 6to4 and for Teredo address, but 0% for IPv4 address.
When running ping6 -n ipv6.google.com and tcpdump -n 'host 192.88.99.1' I see more ICMP6, echo request packets going to 192.88.99.1 than ICMP6, echo reply coming back from 192.88.99.1.
How to debug it? Is 6to4 unreliability probably caused by misconfiguration on my side or problems on some remote side?
Debugging 6to4 problems is almost impossible. Your outgoing 6to4 traffic (to non-6to4 destinations) will go to the IPv4 address 192.88.99.1. There are multiple public 6to4 relays on the internet for that address. You would need to do a traceroute to 192.88.99.1 to find out which one you are currently (it changes over time) using. That is one bit of information. Then it depends which IPv6 destination you are connecting to. They will use the closest 6to4 relay for them. It is almost impossible to find out which relay they are using.
Either of those relays might be causing packet loss.
