Currently using 2 redundant streaming devices that require two separate IP addresses (virtual IP is not an option).
I stream to another server at input RTP x.x.x.1:5004 (streaming device 1, primary) and if that device fails, it fails over to RTP x.x.x.2:5004 (streaming device 2, secondary).
tcpdump shows this happening correctly, x.x.x.1:5004 is hitting my server until I generate a failure. It then starts getting packets on x.x.x.2:5004.
However, as soon as I fail over to x.x.x.2 (or inverse) everything in ffmpeg breaks.
I've isolated this specifically to the ffmpeg process. Running top shows processor % go up 10x what it was prior to fail-over. This is isolated and is the only process happening on this server. RHEL minimal install.