VAAPI on N100 Stops Working After 2 Hours, Requires Reboot

Configuration
1

I am using Frigate as an add-on in Home Assistant. Additionally, I have integrated Double Take, Deepstack, CompreFace, and Plate Recognizer. The system is set up to detect humans as well as vehicles and license plates. Currently, I have only two Reolink cameras connected.

I have been trying for a while to get the GPU to work. I have also tried VAAPI and QSV, but neither worked. It constantly shows 0% usage. The only way I can get it to work is by performing a full reboot of the computer, after which it functions for about two hours.

I have repeated this reboot process three times, and each time the behavior is the same—it works for about two to three hours and then stays at 0% usage without recovering.

I saw that I can try using the i965 environment variable, but I got an error stating that i965 couldn’t initialize. When I disable it, the system works for the two hours I mentioned earlier.

i also open this issue at Github, hope this will bring to end this issue.

2

Hi, i have the same problem with n100. Trying all this methods, and the GPU is still 0%, do you have solution?

3

same issue here.

it used to work in a previous version of home assistant with frigate addon, unsure when this changed but I now observe 100% CPU usage and no GPU used anymore on my N100

4

I have same issue on N150 with HA OS : working for 2 hours, stop working, Hard reboot, working for 2 hours…

GPU Intel UHD 750 MHz GDDR5

5

have you found a solution for the reboot of the Intel HA host PC caused by Frigate HA addon ? I too suffer this problem. It takes a couple of hours and out of nowhere with no log entries my Host PC reboots. (Kernel panic?). I tried everything. CPU load is low (20 to 30%), and I am using an external Google Coral USB for the detector. My Intel is en Atom CPU. (without AVX), I tried both regular version (with or without full access), and the “older intel CPU” version, but they all suffer the same problem. When I look at CPU load, before the reboot the CPU load is frozen to a low value, then recovers after the reboot. I think the freezing takes some 5 minutes and then a reboot is done. Don’t know if reboot is done by HA supervisor or by the CPU itself. I have checked BIOS settings, and also disabled C-states now. But did not solve it. Last what i did now was move the coral usb to a usb2 port instead of a usb3 port, maybe there is too much current drawn. Will see.

6

I have same behaviour as @monty_burns_007. My solution was to disable hwaccel with this config:

# ==== Hardware accel & detectors ====
ffmpeg:
  hwaccel_args: []  # preset-vaapi   # Intel iGPU (QuickSync)

Just commenting out the two lines did not help; I had to include the hard brackets. Also the whole system needs to be rebooted, not only restart the Frigate addon. CPU load increased only by a couple of percentage points. From my understanding I need to wait for a fix in HAOS and/or Frigate to solve the VAAPI / hwaccel challenges rebooting my system regularly.

Update: Item 1 here seems to solve the issue: Frigate 0.15 FAQ · blakeblackshear/frigate · Discussion #16391 · GitHub

1 Like
7

i also read that ffmpeg v7 is broken and you could force to use v5 again to fix this:

fmpeg updates

  • ffmpeg has been updated to version 7. Some Intel users may need to change what hwaccel preset they are using (preset-vaapi may need to now be preset-intel-qsv-h264 or preset-intel-qsv-h265) if camera feeds are not functioning correctly after upgrading. This may need to be adjusted on a per-camera basis. If a qsv preset is not working properly, you may still need to use a preset-vaapi or revert to the previous ffmpeg version as described below.
  • The new version provides many improvements, but the old version can still be run if needed by specifying path: "5.0" in your ffmpeg: config entry. For example:
ffmpeg:
  path: "5.0"