Foreground vs daemon mode, the one-test-at-a-time busy error, multi-port servers, systemd units, and the firewall rules an iPerf3 test server actually needs.
Two tools, not two versions: they do not interoperate, flags collide, and each wins different jobs. Real output from iperf 2.2.1 and 3.18 shows exactly where they differ.
Why fast links benchmark slow: bandwidth-delay product, a real 78% throughput collapse from a 64K window at 104 ms RTT, and how parallel streams buy it back.
UDP mode measures what TCP hides: loss, jitter, and reordering. Real lab captures show a clean test, a 29% loss over-drive, and how to find the loss knee on any path.
Learn iPerf3 from install to JSON automation, with every output block captured live from a routed Cisco lab - including how to actually read Retr, Cwnd, and sender vs receiver lines.