Hello all! I am having an issue with FPP crashing, that I think I may have figured out my own answer, but I'd just like to get the opinions of people who have been doing this longer than I have.
So first some background info. Currently I have a Kulp K16A-B, currently I have just shy of 8,000 pixels connected to it with a combo of the local ports and smart remotes. all 16 local ports have models, and I have a total of 7 remotes attached, 2 of them are technically "dumb" remotes, then i have 2 chains, one with 2 remotes, and one with 3 remotes. The controller also outputs artnet data to control resolume for video playback, and a few other things. This entire controller is running in remote mode since I have 2 pi's running virtual matrix and one pi acting as the player with nothing other than audio attached.
Last year for Christmas everything was running fine, and for Christmas I have just over 9000 pixels attached to it, however I was only using 2 remotes in dumb mode. Now this year for halloween I added more remotes, and I've been having pretty consistent crashes on my controller, tonight there were 11 crash reports, and 2 times where the controller locked up so hard I had to power cycle it.
Tonight I started to attempt to track down the issue, so I took a look at the multi sync page and noticed this particular controller hits 100% CPU and stays there during most of the sequences I have. First I tried to offload the artnet outputs to the player pi since its not doing too much, however that had no impact on the CPU use of this controller. next I thought maybe I had a bad smart remote that was sending junk data back or something, so I pulled all of the remotes (while leaving the configuration) from the unit just to see if it would still crash, it still crashed. I tried putting multi sync into unicast mode thinking maybe multicast was causing problems on my network, even thought I run audio over IP on this same hardware (different vlan) and I know multicast works for that, but it made no change. Finally I thought maybe it was voltage sag on the input or output of the power supplies causing issues so i threw my multimeter on it...while I did learn I should probably run a new circuit for my controller I did not see any issues on the 12v lines powering the BBB, they stuck right around 12.1v. Now i considered bringing out the oscilloscope or spectrum analyzer to see if there was any wild noise that would cause issues, but I think the issue may be easier than that...
So before I go and spend more money on my display this year, would a BBB with a CPU pegged at 100% for most of the sequence cause FPP to crash? Memory seems okay, around 30%, but that CPU number seems like it may be the culprit.
Anyone dealt with anything like this before, and could CPU be causing this problem? Thanks!
TL;DR: BBB is at 100% CPU use with kulp K16A-B when playing sequences, it seems to crash at different places each time during playback, but its only happening during a sequence. Before buying another controller, I wanted to verify if 100% CPU could be the problem.
How are you measuring CPU usage? If using the FPP About page, IMO, it is not very reliable as it tends to spike when the About page loads and almost always reports 100% CPU usage. Log into the instance using the ssh shell and use TOP to view cpu usage. TOP will also break down the running processes and you can see which processes are using CPU resources. Here is a screenshot of my idle FPP player instance (K8 with BBB).


