This isn’t necessarily a Plex specific problem I don’t believe, but the fix was to disable it’s nightly maintenance scripts.
By default, several scripts kick off at 2am, and the behavior seems to be that on a very large media server (30+TB), running ZFS, even with ARC cache limited, the scripts would take so long that FreeBSD’s default nightly (and weekly) maintenance routines would launch at 3am, and 4am respectively, backlogging them and something happening where if they’re all running at the same time, the OS would get a kernel panic and reboot. It was pretty consistently rebooting either nightly, or every 3-4 days at most. After disabling Plex’s nightly routines, it’s been fine since. I’m also taking up the topic with the FBSD dev team as well, but thought I’d put this out there for anyone else that may be experiencing it.
