mctr fails to accept all HC connections

Summary

If a large number of HC try to connect to the mctr, some of them are rejected.

Steps and/or TTCN-3 code to reproduce

Try to start ~ 30 HCs at the same time.

What is the current bug behavior?

Some HCs can't connect, because the TCP connection is refused

What is the expected correct behavior?

All HC connects

Possible fixes

When the mctr opens the listening port, it sets the backlog to 10. When a new connection is accepted, a possible time consuming reverse DNS query is executed. If the number of the waiting connection exceeds 10, the new attempts are rejected by the kernel.

simply increase the backlog. The current value remained unchanged since the beginning.

"If the backlog argument is greater than the value in /proc/sys/net/core/somaxconn, then it is silently capped to that value. Since Linux 5.4, the default in this file is 4096; in earlier kernels, the default value is 128."

So it is safe to increase it to 4096

Titan version

all

Platform details (OS type and version)

all

/cc @aknappqwt