Server Version#: 1.14.1.5488
Player Version#: 3.85.2
(Linux Mint 19.1)
For some time now, I’ve noticed that in the middle of the night when I have PMS set to do it’s scheduled tasks, the thread for Plex Media Server will ping at about 50%, even when no other subprocess is running, and stay that way indefinitely until I restart the server. Scheduled tasks are set to run at 2:00 a.m. and stop by 8:00 a.m. All the little things I found in similar posts that could spike the CPU, like updating all libraries during maintenance, and upgrading media analysis during maintenance, have been turned off. Database is optimized. Library scans are set for every 12 hours and partial scans are run when changes are detected, but “Scan my library automatically” is off. Video preview and chapter thumbnails are off, as is “Enable video preview thumbnails” for every directory in the library. Trailers and extras are turned off.
This will happen whether anything has been added to the server or not. I’ve let the process run to see if/when it will stop, and it has stayed at around 50% for close to 24 hours before I restarted the server.
I’ve checked other posts on this problem, and have seen a number related to VM’s, but this installation isn’t on a VM. Also saw this post which had a 100% spike related to UPNP; I tried turning of UPNP, but didn’t notice any change, and can’t turn that off in my router anyway because we use Chromecast with Plex, and UPNP is required for Chromecast to work. (Besides, I’m only pinging out at 50%, not 100%.)
Been using PMS for years, but this is my first post. Have logs if needed, but not sure which ones to attach, so any information would be appreciated.
/ Hey, I turned off the extensive media analysis during maintenance, and this is the first night the CPU hasn’t spiked and stuck in I don’t know how long. So it looks like that was the culprit.
False alarm. It seemed fine last night – no endless spikes after turning off “extensive media analysis during maintenance.” Used PMS to watch two half-hour shows earlier tonight and listen to some music (no transcoding), went to my machine a couple hours later when it wasn’t in use, and the CPU was pinging again: plexmediaserver at 50% CPU with no subprocesses running, PMS not being used (for hours), and the web page not even open. This was also 3 hours before it was meant to perform scheduled tasks; in the past, that spike seemed to occur during that scheduled tasks period and not stop. Now it occurred well before the scheduled tasks period.
So it doesn’t seem “extensive media analysis during maintenance” wasn’t the culprit. Restarting the server again.
Turned off “Refresh metadata periodically,” as well as “Perform refresh of program guide data” since I’m not using the EPG (I just use PMS for really basic things). Will see if this makes a difference. If not, I’ll turn off all the rest of the scheduled tasks and see what happens, but that still leaves the question of why it pinged 3 hours before the scheduled tasks began.
Don’t jinx it lol