Intermittent Issue: NTTTCP on server side does not exit after test completes

Hi Team,

I noticed an intermittent issue while using ntttcp-for-linux for one of my benchmark tests.

Requirement

I wanted to collect TCP throughput for 192 cores VM. I wanted to run ntttcp for 10 iterations. Each Iteration is a test of 60 seconds with 10 seconds warmup.

Following are the commands used:
Server command: ntttcp -r -p 7101 -t 60 -m 1,*,10.1.0.4 -b 1M -W 10 --show-nic-packets eth0 --show-tcp-retrans
Client command: ntttcp -s 10.1.0.4 -p 7101 -t 60 -m 1,*,10.1.0.4 -b 1M -W 10 --show-nic-packets eth0 --show-tcp-retrans

Issue

I noticed that sometimes (intermittently) that server side ntttcp got hung after the test execution time and publishing the server result and didn't return to the command line. The client side ntttcp exited properly. I had to perform Ctrl+C to exit NTTTCP and recover the command line on the server VM.

Though very intermittent, I have seen the issue with other combinations such as 32 threads and 64 threads as well. Issue is noticed for both TCP and UDP protocol tests, majorly on VMs with higher number of cores.

No source file, test, or entry point is named. Start by reproducing the paired server and client commands from the report, varying TCP/UDP and thread counts on higher-core VMs, then compare completion behavior. Done means the server publishes its result and returns to the command line without requiring Ctrl+C.