Fan is quite loud since the last update
Edit: bug solved
Hi I have a Minisforum um 780 xtx , it feels like the fan is pretty loud since the last update, and the CPU clock rate doesn't go down, I also have a higher CPU load in idle. Before it was at 400mhz (Even if KODi was running). And you could hardly hear the fan at all. Now the CPU is average at 2.4ghz
23 Replies
I believe this is because mangoapp keeps crashing and restarting, as my mini pc cpu fan keeps spinning up and down whilst the system is supposed to be idle, which is unusual. I think the bazzite devs are aware of this issue.
That actually sounds like a cause, thanks, I hope it gets fixed quickly, mine is actually on 24/7 and is in the bedroom, the fan noise is pretty annoying
I have created a bug report on github in relation to this. https://github.com/ublue-os/bazzite/issues/2275
GitHub
mangoapp constantly crash/looping and core-dumping - Causing spikes...
Describe the bug Hi, Am using bazzite-deck-gnome:stable on an AMD 7840HS Mini PC and have noticed that since the last 2 stable updates that at idle in gaming mode, the CPU fan constantly spins up a...
Same issue with mangoapp crashing.
journald: https://paste.centos.org/view/9160a4a7
rpm-ostree status: https://paste.centos.org/view/bd78b137
@adr3nal1n Can you try
bazzite-rollback-helper rebase 41.20250131 and post your results?
That build is the last one without the crashing issue for me.
if you wanna go back to latest stable you do bazzite-rollback-helper rebase stable
logs for 41.20250131:
https://paste.centos.org/view/fe90d7f0
No more crash loopI have just downloaded the latest update, same behavior, fan and CPU revs up. Now I'm going to downgrade Will report soon
It worked, thank you very much, CPU clock speed goes down again and the fan is barely audible. I hope the crash logs haven't put too much strain on my SSD in the last few days
nah you're good
its not been an issue for a long time
Five days 24/7 every three seconds 🥲
how big is the SSD? if its like 1TB+ it's not too bad. you can always check how much total TB written by using smartctl
sudo smartctl -a /dev/<devicename>
it will be called something like "data units written" or "Total TBW written" or something similar
it depends on the ssd
you can always check with device manufacturer how much TBW the ssd supports
and compare
usually its 600TBW on 1TB
1200TBW on 2TB etc
a few days thus does not matter19.619.485 [10TB] It's only a few months old and it only runs KODi / IPTV
nothing to worry about
i think i agree that maybe its a good idea to disable storage of coredumps for repeating process crashes. that it something that systemd should catch and control, and not do this. It's not really the fault of bazzite. mangoapp should also not be forced to be re-opened after repeated process crashes. that i believe is just how valve implemented it, but im not sure
its just a massive chain of shit that lead to tremendous ssd write amplification
Issue resolved after updating to latest stable release! Thank you so much Bazzite devs!
Yep it is resolved for me too
@RICK_C Update to latest stable resolves the issue. Just a tag so you're aware its fixed 🙂
I’m using the Nvidia-deck image on a htpc and I’m still experiencing the issue. Just updated to the newest stable release and the clock rate and temps are still high while idling in game mode menu. 75+ degrees Celsius average while idle meanwhile the cpu temp is in the 60s in a game like stalker 2 even during shader compilation. Should mention that the temp used to be way less before when idling in game mode
post logs:
ujust logs-this-boot | fpaste
if its the same issue, it will be very easy to tell
also make sure you're 100% sure you're on the latest version with ujust update
By the way: Do not use Game Mode to update Bazzite. There's currently a bug that's making it not install the latest update properly. Use the above command in terminal.https://paste.centos.org/view/5611eaab
Oh and the image is bazzite-deck-nvidia-gnome. Forgot to mention it was the Gnome version sorry about that
Not the same issue. This issue was about mangoapp crash dumping over and over again, which was eating up cpu cycles and ssd write amplification. It is not GPU related.
Your logs are clean. This is some other thermal issue with your nvidia card that requires more digging into what is causing high idle usage.
You will need to monitor GPU usage and power states and figure it out that way. Often this also means the CPU is being used (and possibly high RAM fluctuations).
It could also be that you're using something like a 240Hz monitor, which will naturally increase idle usage when screen is on.
Sorry I just realized I made it sound like it was my gpu. It’s my cpu idling at 75 degrees in the game mode menu
The gpu is like in the 30s in the game mode menu
It’s weird cause last week the cpu was in the 30-40s
And the clock rate doesn’t go down at all
Could be some kind of regression, but I have no clue since this is not the same issue
4% cpu usage is not a lot
you have changed nothing hardware wise over the past weeks?
You could add the system manager as a no Steam app, then you can start it in game mode and see which process is using the CPU
Just checked the machine and I’m an idiot. Somehow I knocked out the fan connector of the fan that’s directly under the cpu cooler (mini-itx build).
Seems to be running much better now at 50 degrees
That made the most sense to me. I saw 4% usage at almost 70c. Made 0 sense to me, even if the frequency is kept at maximum.
However, I'm not sure it's normal that the cpu multiplier doesn't go down while idle?
Have you disabled AMD Cool 'n' Quiet in BIOS? IIRC That is what determines whether the CPU will lower its multiplier while idle.
That's why I would have guessed that some process was loading the CPU, 4ghz in idle is unusual For me it goes down to 400hz by default