We are removing the hcimonitord/pid and cmd_port files.
Initially we did not realize this was occurring sometimes and so we ended up with 2 hcimonitord daemons running.
today I had the issue, luckily on a non-prod site. We finally got it to shutdown and stay down (so I can do site cleanup) after turning off all the non prod sites.
I am cycling the hcimonitord logs weekly via a script.
any suggestions are appreciated