yap yap ... tcp in tcp is really bad (learned it while setting up a vpn with ppp-over-ssh 10 years ago ;)).
tinc via G3 is working fine for me after i did some experiments - don't know if it works for you too. i used a server in public net to connect all clients (only 2 + 1 server in my case) with tcponly-option. this server has no services i need in the vpn - so it is only used for exchanging meta-information. than i added a connectto option for each client on both sides, the server that have to be reached and the client it self both WITHOUT the tcponly option. this should result in a kind of firewall-hole-punching technique. another possible cause for the disconnect: if your provider uses Nat it have to keep a Nat-table. tinc only opens one connection and keeps it open. if the Nat-router have a timeout for tcp-connection it will be dropped, silently. but this should be solved to with the upper setup. please let us know if you got it working ;) regards sascha On Thu, 2009-03-26 at 18:08 +0100, Guus Sliepen wrote: > This is degrading the performance. Also, the kernel will > buffer packets of TCP connections of both tinc and applications, > causing very > long delays. _______________________________________________ tinc mailing list [email protected] http://www.tinc-vpn.org/cgi-bin/mailman/listinfo/tinc
