Hi all, I'm having an issue where I have high CPU usage when the server is basically idling. The worst offender is core 2, which pegs at 95% when this problem crops up, but I can see that the other 5 cores also jump up by 20% as well. This can make the server unresponsive in webui and I've been having random connection drops in Plex which I assume is the same problem. 

 

I noticed that the syslog was full and when I checked in htop I saw that the rsyslogd process seemed to be consuming a lot even though the syslog server is disabled in settings. I killed it but the problem persists. 

 

This started after I upgraded my server into a new PC so perhaps it stems from the hardware change. I attached the diagnostics, when I looked through the syslog in it I noticed that it had a ton of PCIE bus errors which must have been why the log was full. I have rebooted since pulling these diagnostics and the issue is still present. Checking htop again I see that the high usage I see on the dashboard page is matching when rsyslogd is consuming a lot of cpu. (Interestingly there's 2 processes of it in there)

 

Help pls??

gigamulch-diagnostics-20250516-1547.zip

I also feel like it's maybe doing more network traffic than the docker containers account for... Is there any reasonable chance someone could be accessing my server and running things in the background without me knowing?

I'm in the logs and I see it's having an error again and again about the Marvell 9215, maybe it's working the CPU to death trying to correct that, and the syslog process was also busy writing the issues down. That could explain why the issue isn't resolved when I close that task. I don't really wanna load the parts cannon at it but I'm seeing discussion about these being problematic cards

There are constant PCI errors logged, see if this help:

https://forums.unraid.net/topic/118286-nvme-drives-throwing-errors-filling-logs-instantly-how-to-resolve/?do=findComment&comment=1165009

I pulled the sata expansion card and it seems to be happy. I assume it's just because it's a marvell controller that it has issues, and that other things plugged in to that slot won't have problems. I didn't bother messing with aspm off as the thread you posted suggested as I actually had enough sata ports without it, and I really like my server sipping power when possible. Now to find a sata expansion or HBA that is reliable and doesn't block C states. As is my power draw is down by almost 20w. Thanks Jorge!

Ugh my Other sata card is also throwing occasional errors, maybe 1-2 a day. I'll probably ignore them for now seeing as it is a JMB582 controller which is on the recommended controllers list.