I recently build a nas system and installed the latest version of Rockstor on it. Right away I noticed from the web UI that the processor usage never drops below 25% (Intel core i3-6100). After SSH’ing into the NAS and running top I see that kworker/0:1 is using roughly 100% of one thread of this CPU. As this system is pretty new I hoped that the process would actually idle out after some time, but after running the nas for a couple of weeks nothing had changed. Also, rebooting didn’t resolve my problem.
The system I’m running is quite simple. The system boots from an USB 3.0 32GB stick and only one 4 TB drive is used for file storage. I do run a couple of rock-ons, namely plex, plexpy, nzbget, sonarr and couchpotato. However, shutting down the rock-on service doesn’t resolve the problem. Does anyone knows what causes this or a way of running deeper diagnostics? Hopefully, you can help me resolve this problem.
This indicates the kernel is stuck performing some of its (many) tasks and as such has a range of possible causes. The best thing to do is to yum install perf and follow steps 2 and 3 in this post to see which module is behaving erratically.
Thanks for you quick response. I’ve installed perf and ran the commands mentioned in the post. A lot of rows appeared in the results. I’ve attached a screenshot of the output shown in the terminal below. Could you help me make sense? Thank you very much for your help so far.
Was kworker/0:2 the erratic thread in this case? I forgot to mention that for details about a particular process you can use the -p <pid> option instead of -a, e.g. perf record -g -p $(pgrep -x kworker/0:2) sleep 10.
However, in the results you posted the kernel seems to be busy with a lot of ACPI calls. You could try this post for further diagnosis/workarounds.
