Matthieu Baerts
Matthieu Baerts
> Adding answers inline: Thanks! > * **In your setup with 4 endpoints on each side, which paths do you expect to have?** > overall, the goal is to have...
> unfortunately, not an option in our case to "add specific routes" or control application in any way. In this case it's a set of ephemeral machines with multiple network...
> So far i tried latest kernel 6.17 and also laminar endpoints (merged mptcp_net-next on top of "vanilla" 6.17), but in both cases still see the same behavior. Maybe I'm...
> good catch, that looks like a reason, on re-run it was using different server endpoints (even without latest 6.17 kernel nor laminar endpoints). > > Although, i noticed that...
> yes, right, it was a test on older kernel, i tried to configure laminar endpoints as you described and it evenly distributed subflows (iperf with 128 connections and 4...
> Tried to start listening sockets in background for all endpoints on "iperf3 server" side, but no luck. It's still stuck to a single server IP E.g. I just tried...
Hello, I did some investigations: with IPv6 link-local addresses, the network interface needs to be specified upfront. Did you then modify your Python script to call `SO_BINDTODEVICE`? s.setsockopt(socket.SOL_SOCKET, socket.SO_BINDTODEVICE, str(sys.argv[2]...
If it helps, you can look at the new tests I just sent to validate cases where the listener 'bind()' to a specific IP address, and other listen sockets are...
Hi Gang, Thank you for looking into this issue. So, from what I understand, it is still useful, and it shouldn't be removed. Is that correct? If yes, should we...
Hi @Dwyane-Yan, > I think this variable is currently not useful If this variable is currently not useful -- e.g. to avoid taking lock in the fast path, stopping a...