Server Version#: 1.15.0.659
Player Version#: N/A
Just built a new PMS with an intel i7 8700K and have a constant 10% usage on Task Manager with no activities at all. Because of that level of activity my cpu stays at 4.5ghz causing extra noise from the fans. I thought I fixed the problem yesterday by turning off any extra features in settings and after having no effect I turned them all back on and after a reboot the PMS activity went to near zero and my cpu throttled down to 800mhz. I even tried playing several demanding videos and after I stopped again cpu throttled back to 800mhz. Unfortunately after a reboot today the constant 10% is back and repeatedly trying what I did yesterday I can’t fix the problem. Any ideas on what I can do to resolve this issue? Thanks.
Things to try:
- Disable DLNA in Plex
- disable UPnP in your router
Restart Plex after changing each
Tried turning off both DLNA and UPnP rebooting server each time with no change. I did check CPU activity early today and saw the Plex Media Server only using 0.2% for several hours but as soon as the computer needed to be rebooted the activity went back to 10%.
How about letting it just sit and see if the activity eventually dies down?
This activity has been at 10% non-stop for a week. If I shut down PMS my cpu usage immediately drops to less than 2%, cpu freq drops to 800mhz and my power usage drops 20 watts. Also it seems my hard drives never shutdown with PMS running, drive temps stay above 45. I have been a huge supporter of Plex for years but in frustration installed Emby for a day and didn’t have any of these issues. Emby also gives you complete control over hw transcoding allowing me to choose my i7’s iGP over my nVidia GPU (something Plex doesn’t allow). Should I fill out a bug report or is this just normal behavior for PMS and I should move on to something else?
Is there something special about your media storage?
Is it locally in the server or elsewhere?
All of my media is stored in 4 usb drives plugged into the computer with PMS running.
Do you have several router-like devices on your local network?
I have a lot of devices on my lan. I have 1 router as the primary router and a second one on the other side of the house setup as an access point to extend my wireless coverage. I shutdown the second router for an hour and saw no difference cpu usage of PMS.
Can you enable debug logging
- activate debug logging (not ‘verbose’!)
- quit Plex Server
- wait 1 minute
- start Plex Server
- wait until the issue starts again (it’d be very helpful if there were no other activity on your Plex)
- wait 3 minutes
- fetch log files and attach them here
Your comment about my lan sparked my curiousity. I disabled my lan adapter to make sure there was nothing on my network causing the problem and just as I expected cpu activity remained high. As a double check I left the lan adapter disabled shutdown PMS and then restarted it. The high cpu activity was gone! I re-enabled the lan adapter and immediately it returned. I watched my Network activity and it remained at zero. WTF! I don’t know what is going on but at least I have an avenue to pursue. Any ideas? Do you still want me to send you Debug Logs?
Thanks
Plex Media Server Logs_2019-03-01_16-45-25.zip (5.2 MB)
Requested logs.
Was this lan adapter the only network adapter in your machine? Or do you have several of them?
The motherboard I am using is a MSI Mag Z390 Tomahawk and yes it does have two Lan adapters, an Intel 1219-V and an intel 1211. I tried both and even disabled the unused adapter and didn’t see any change.
Our expert is asking for another 2 sets of log files and accompanying process memory dumps.
1st set:
- enable ‘verbose’ logging and restart Plex server
- wait until sustained 10% cpu load + then create a memory dump using the windows task manager (CTRL+Shift+Esc - processes, right-click on the process which creates the load - ‘create dump file’)
- fetch the log files
2nd set:
- disable ‘verbose’ logging (but keep ‘debug logging’ activated)
- restart Plex server
- wait for sustained cpu load and produce second dmp file and gather second set of logs
Upload both sets of dumps and logs please.
Hi sigfreund,
I’m also have a problem with 1.14 and 1.15. After I have downgraded to 1.13 in Centos the CPU usage is gone.
Can you try this version?
I think the CPU usage is high because of Dashboard with usage perfomance in Plex Web Gui.
Something I didn’t mention earlier because I removed it from the system and it didn’t make a difference is a whole home security box called Bitdefender Box 2. As more and more of your suggestions seemed to be network related I eliminated everything possible and started with my Plex server connected directly to my cable modem. After a reboot the excessive cpu usage was gone. Then I put the Bitdefender Box 2 in between the cable modem and my Plex server and the excessive cpu usage was back. I have had the Bitdefender Box 2 in my system for years and am sure it wasn’t causing a problem before so the next step is to try downgrading version 1.13. Should I send logs with and without Bitdefender Box 2 in the loop and before and after downgrading to 1.13?
If you are ready to invest the time, yes please!
Every data point is valuable for diagnosis.
Please don’t forget the memory dumps.
Here are the logs for version 1.15Plex Media Server Logs_2019-03-02_13-13-25-NoBitdefenderBox.zip (5.0 MB)
Plex Media Server Logs_2019-03-02_13-47-12-WithBitdefenderBox.zip (5.3 MB)
Downgraded to version 1.13 and excessive cpu usage is gone. Here is the log.Plex Media Server Logs_2019-03-02_13-52-04-V1.13.9.zip (4.9 MB)
