-
Notifications
You must be signed in to change notification settings - Fork 73
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
NVME U.2 drive temperature not considered for fan duty cycle #412
Comments
Update: I said earlier maybe the case fans and CPU fans should be set independently. I don't think so anymore. All of them on MAX still aren't enough to cool the U.2 drives. (Actually just yesterday I bought and installed 2 more U.2 drives, labeled "Intel1" and "Intel2" in the NVME Temperature plot below. I didn't include them in the plots above because they weren't installed until late in the day Aug 8th and didn't want to add confusion. But including now for completeness.) You can see just before 11:00 AM I modified my fan override script to also set the CPU fan to MAX:
Looks like the temps on the SSDs are stabilizing... above 70° for 2 out of the 3 U.2 SSD drives. At least the GPU fan doesn't have to work as hard... |
I figured out how to get the HDD temperature from smartctl. The lifetime max temperature for /dev/sda is 72° C! (Even within the last few days it has hit 71° C.) The maximum recommend temperature (according to smartctl) is 55° C:
/dev/sda is the "Seagate Barracuda 2.5 5400" drive (ST5000LM000-2AN170) that came with the machine. |
Distribution (run
cat /etc/os-release
):Related Application and/or Package Version (run
apt policy $PACKAGE NAME
):I'm using this patch as well, as my computer is not stable without it: #321
Issue/Bug Description:
I happened to notice that my NVME U.2 drive is running really hot (80° C). According to https://superuser.com/questions/1592187/should-i-worry-about-high-ssd-temperature 70° C should be the maximum operating temperature.
Here is the current smartctl output (note the FAILED status due to temperature):
Steps to reproduce (if you know):
Do some heavy I/O tasks and watch the drive temperature increase. As long as the CPU load isn't too high the fan speed will NOT increase.
Expected behavior:
Case fan speed should increase when U.2 drive temperature (and/or 2.5" HDD drives!) increases.
Other Notes:
Here are some pictures generated from my logs (I have a custom script to record temperatures every 5 seconds):
The NVME drive labeled "Samsung" is the boot drive (stick on the motherboard). The drive labeled "Micron" is the 2.5" U.2 drive. I wish I was recording my 2.5" HDDs, but I'm not. (Daily cron jobs run at 4:40 AM everyday which restarts the monitor script (and other cron jobs), hence the spike in temps and sudden end of the log.)
Note this is the CPU fan command, not the case fan command. But all fans are set to the same duty cycle so it's a fair proxy.
Note the 2 sets of FAN RPMs. The higher one is the CPU fan. The lower one is the 2 case fans.
In my opinion the U.2 (and 2.5" HDDs) drive(s) temperature should also be considered in calculating the case fan speed. Right now the logic is to use the CPU temperature (and NVIDIA temperature, if available) to set both the CPU fans and case fans.
My plan is to write a patch to do so, as right now my U.2 drive is slowing to a crawl and btrfs is throwing errors in syslog due to the high temps.
For now I wrote a script to force the 2 case fans (label INTF in /sys/class/hwmon) to max speed (trying to override what system76-power is doing):
And it does seem to help (you can see I started it at about 10:15), but the drive is still running hot (but slowly coming down):
The text was updated successfully, but these errors were encountered: