On dim. 26 mars 12:49:44 2017, Andrew Savchenko wrote: > Run tcpdump -w on both sides. Compare dumps when connection stalls > and when it works fine. Many reasons are possible, it's hard to > guess from data you provided.
I did this, I saw many neighbor solicitation on pokedex’s side (the lossy side) when the loss happens. I’ve also included some monitoring packets, but they didn’t match the loss window time. However, they are flapping accordingly to the loss in the monitoring web interface. https://bulbizarre.swordarmor.fr/garbage/documents/argall.pcap https://bulbizarre.swordarmor.fr/garbage/documents/pokedex.pcap > But it makes me wonder why you have default via VPN and given > address via eth0. This may lead to undesirable consequences like > VPN carrier (or some aux request) trying to go through its own VPN > tunnel. The only goal is to reach the server from its two public addresses. As each provider is doing BCP38, I have to set up multiple routing tables to make the traffic flows to the correct interface based on the source address. > Best regards, > Andrew Savchenko Regards, -- alarig
signature.asc
Description: PGP signature
