Dear Manjaro community
I am at a loss...I need your help as I am too new to this.
About a week ago I changed from Windows 10 to Manjaro 20.0.3 (Xfce).
Everything worked fine. I had no troubles. No freezes, no crashes, it worked really good.
Until yesterday evening...when I started up this morning I got freezes all over the place.
It seems that some kernel thread is hogging the CPU making the freezes, but I can't remember changing my kernel (even tho I happen to be on 5.6.16 now instead of 5.6.15 all of a sudden).
How can I change that back to check it?
Until yesterday evening everything ran smoothly and fast (for about a week now). All applications were rather quick to start and use.
Today I booted up and had freezes (of about 3-5 seconds) whenever I started an application, when I clicked the "menue" button of Xfce, when I opened something up, switching applications, etc.
The whole screen froze it seems, funny enough youtube kept the sound playing, only the video seemed to freeze and continued (at the time where those 5 second passed).
I have no changed anything on the hardware (no new or changed hardware, no firmware updates, etc.). I even didn't install new software or updated software yesterday (at least not that I know of!).
AMD Ryzen 9 3900X on a Asrock Taichi X570 and a Sapphire 5700XT Nitro+
32 GByte DDR4-RAM
No overclocking on GPU, CPU or RAMq
Two NVMe SSDs M.2
One SSD on SATA
Manjaro is on the "primary" NVMe M.2 (1 Tbyte)
Manjaro 20.0.3 Xfce installing medium used, updated the last time two or three days ago (not yesterday, that I know) and I am using kernel: 5.6.16-1-MANJARO #1 SMP PREEMPT Wed Jun 3 14:26:28 UTC 2020 x86_64 GNU/Linux
HOWEVER: as far as I can remember, I was on 5.6.15 (and not on 5.6.16), but I CANNOT remember installing a new kernel
Diagnostics so far:
I tried to use iotop, htop, nvtop and some more, but the freezing didnt allow me to see what happens when it happend (it was normal after such a freeze).
So I tried to ssh into my box from my laptop, hoping I can cause the freeze using my keyboard and mouse while checking all the monitoring tools from ssh...that worked.
sensors gave me less than 55 degree celcius on all the sensors I had (CPU, GPU and nvme), so I doubt its an overheating issue
When a freeze happens one CPU core spikes to 100% during a freeze (htop red colour - meaning kernel thread) - and it seems its always the process of Xorg that goes berskers with >99%
(The exact process is /usr/lib/Xorg :0 -seat seat0 -auth /run/lightdm/root/:0 -nolisten tcp vt7 novtswitch)
RAM was used to about 3 GByte (a tenth). No spikes there, no swap use
iotop was OK as well (no spikes), the processes that showed up with -Pao where all >0.20%, so I gather its not a I/O issue)
Anyone an idea what else I can check or know of such symptoms?
Why did it start out of the blue?
What would be the tool to use to get back to 5.6.15 to check out if there is a regression (sorry, I am new to manjaro)
I just checked the manjaro settings manager (and the CLI command) and there is no option to "downgrade" as far as I can see.
Indeed there was 5.6.15 first which was updated some days ago in the repos to 5.6.16. I am not aware if and when this is updated automatically...considering the kernel thread CPU spikes I'd like to get back to 5.6.15 to test out if there is a regression.
However, for the life of me, I cant find a tutorial on google search or forum search concerning "downgrades" of current stable kernels (my google/search fu is bad today, sorry)