frigate: [Support]: Frigate 0.12.0 randomly crashes

Describe the problem you are having

Experienced twice during the past two days. The docker container seem not responding. If I try to stop it, I get an error message saying that the container cannot be stopped (“tried to kill the container, but did not receive an exit event”). I see ‘go2rtc-healthcheck service exited with code 256 (by signal 15)’. Then it says ‘service go2rtc-healthcheck successfully stopped’. I did not have this issue when running 0.12.0-rc1 (I had rc1 running for about two weeks before started using the official 0.12.0 release yesterday).

Version

0.12.0-DA3E197

Frigate config file

# I have 6 cameras setting up the same way.  Only present settings for one here
mqtt:
  host: 192.168.xxx.xxx

detectors:
  ov:
    type: openvino
    device: AUTO
    model:
      path: /openvino-model/ssdlite_mobilenet_v2.xml

model:
  width: 300
  height: 300
  input_tensor: nhwc
  input_pixel_format: bgr
  labelmap_path: /openvino-model/coco_91cl_bkgr.txt

birdseye:
  enabled: True 
  mode: continuous

ffmpeg:
  hwaccel_args: preset-vaapi
  output_args:
    record: preset-record-generic-audio-aac

go2rtc:
  streams:
    front_door:
      - rtsp://xxxx:xxxx@192.168.xxx.xxx/live
      - ffmpeg:front_door#audio=aac

  webrtc:
    candidates:
      - 192.168.xxx.xxxx:8555
      - stun:8555

detect:
  enabled: True
  width: 1920
  height: 1080

rtmp:
  enabled: True

cameras:
  front_door:
    ffmpeg:
      inputs:
        - path: rtsp://xxxx:xxxx@127.0.0.1:8554/front_door
          input_args: preset-rtsp-restream
          roles:
            - detect
            - record
            - rtmp
      output_args:
        record: preset-record-generic-audio-copy

Relevant log output

The Web UI crashed, I have no way to get any log.

FFprobe output from your camera

Container crashed.  I cannon access Frigate Web UI, or open a shell to access the container.

Frigate stats

Frigate Web UI crashed.  Cannot reach the end point.

Operating system

Other Linux

Install method

Docker CLI

Coral version

Other

Network connection

Wireless

Camera make and model

Wyze Cam V2, V3 on rtsp firmware

Any other information that may be helpful

The application was run on a QNAP NAS with J3455 CPU. This is the command to create the container: docker run --name=frigate --shm-size=256m --restart=unless-stopped --env=TZ=America/New_York --env=LIBVA_DRIVER_NAME=i965 --env=FRIGATE_RTSP_PASSWORD=xxxxxx --volume=/share/Container/frigate/config:/config:rw --volume=/share/share_vol2/frigate/media:/media/frigate:rw --network=bridge --privileged --workdir=/opt/frigate -p 4xxxx:1935 -p 4xxxx:5000 -p 4xxxx:8554 -p 4xxxx:8555 -p 4xxxx:8555/udp --label='com.qnap.qcs.network.mode=nat' --label='com.qnap.qcs.gpu=False' --memory="4g" --cpus="2" --detach=true -t ghcr.io/blakeblackshear/frigate:0.12.0 I used the same command to create 0.12.0-rc1 container. Config files are identical in 0.12.0 and 0.12.0-rc1.

I know it is hard to debug without any logs. I recall there was one object referencing from one of the cameras stayed at ‘in progress’ stage for very long time. Not sure if this exhausted system resources. But I would assume same thing also happened in 0.12.0-rc1, which never crashed. Another observation is that, I ran the docker from QNAP’s Container Station (similar thing to Portainer). When Frigate crashed, Container Station lost control of it. I had to turn off the NAS by pressing the power button. There seems to be a hanging process. When I run the ps command, the command just hangs there after showing some processes.

About this issue

  • Original URL
  • State: closed
  • Created a year ago
  • Comments: 37 (18 by maintainers)

Most upvoted comments

Histogram doesn’t tell us much besides stating there was high-usage. As @NickM-27 mentioned, showing an htop of the processes is necessary to identify which processes is consuming the resources.

wyze docker bridge connects locally to the camera, just FYI @idle-user

Thanks. Are you using 0.12.0 official release and go2rtc restreaming?

Yes.

How much RAM do you have? My 0.12.0-rc1 with go2rtc config crashed as well after around 10-11 hr running. So I now have total of 3 crashes when using go2rtc. I realized that my previous try of rc1, which lasted about two weeks, was using a config without go2rtc and restreaming. I will revert to non-go2rtc and see.

The DS1520+ has 8GB of RAM.

My system overloads and runs out of memory within an hour. Since I have managed to stop it, I assume I avoid the crash you’re facing?

I have a strong feeling it has to do with an RTSP Wyze camera I have. Ever since I have disabled the camera in the config, I have been running fine since yesterday.

Still monitoring