[Support]: Frigate partly crashes after around 24 hours of uptime #19859

BlueTree242 · 2025-09-01T13:20:41Z

BlueTree242
Sep 1, 2025

Checklist

I have updated to the latest available Frigate version.
I have cleared the cache of my browser.
I have tried a different browser to see if it is related to my browser.
I have tried reproducing the issue in incognito mode to rule out problems with any third party extensions or plugins I have installed.
I have asked the AI at https://docs.frigate.video about my issue.

Describe the problem you are having

After having Frigate running, after around 24 hours, I login to frigate to checkup on the camera, from the main screen I can see the "No frames have been received, check error logs" image instead of camera snapshot. Additionally, live view works perfectly well, showing that there is no issue with the camera or WebRTC. The recordings also stop the moment this "crash" happens. This also happens with an error:

2025-09-01 06:01:35.216773317  Exception in thread Thread-9:
2025-09-01 06:01:35.216775950  Traceback (most recent call last):
2025-09-01 06:01:35.216777173    File "/usr/lib/python3.11/threading.py", line 1038, in _bootstrap_inner
2025-09-01 06:01:35.216778008      self.run()
2025-09-01 06:01:35.216779023    File "/opt/frigate/frigate/video.py", line 244, in run
2025-09-01 06:01:35.216782003      while not self.stop_event.wait(self.sleeptime):
2025-09-01 06:01:35.216783866                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2025-09-01 06:01:35.216785454    File "/usr/lib/python3.11/multiprocessing/synchronize.py", line 349, in wait
2025-09-01 06:01:35.216811250      self._cond.wait(timeout)
2025-09-01 06:01:35.216812798    File "/usr/lib/python3.11/multiprocessing/synchronize.py", line 252, in wait
2025-09-01 06:01:35.216813975      self._sleeping_count.release()
2025-09-01 06:01:35.216815139  ValueError: semaphore or lock released too many times

This is followed by an infinite spam of:

2025-09-01 06:02:54.725964321  [2025-09-01 06:02:54] frigate.record.maintainer      WARNING : Too many unprocessed recording segments in cache for camera_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest...

Since I experienced this issue 4 times already, I already had debug enabled, the following was printed all the time until the thread crash and was never printed again since:

2025-09-01 05:59:04.749399789  [2025-09-01 05:59:04] frigate.record.maintainer      DEBUG   : Copied /media/frigate/recordings/2025-09-01/05/cam/58.51.mp4 in 0.019021987915039062 seconds.

obviously sometime around 100ms every time, not this exact number every time

So, my storage is more than fast enough to keep up with my recordings.

Steps to reproduce

Configure frigate & go2rtc.
Leave it on for around 24 hours and check on it.
Observe "No frames received, check error logs", while live view works perfectly.

(I am assuming it is not that simple since obviously there are thousands of people already running frigate with no issues, but I have no idea what triggers this).

Version

0.16.0-c2f8de9

In which browser(s) are you experiencing the issue with?

Google Chrome

Frigate config file

# MQTT Configuration (Disabled)
mqtt:
  enabled: false

go2rtc:
  streams:
    cam_name:
      - rtsp://admin:[email protected]:554/h264_stream#timeout=2s

ffmpeg:
  hwaccel_args: preset-vaapi

cameras:
  ezviz_cam:
    ffmpeg:
      inputs:
        - path: rtsp://127.0.0.1:8554/cam_name?video&audio
          roles:
            - detect
            - record
            - audio
      output_args:
        record:
          - -f
          - segment
          - -segment_time
          - '10'
          - -segment_format
          - mp4
          - -reset_timestamps
          - '1'
          #- -use_wallclock_as_timestamps
          #- '1'
          - -fflags
          - '+discardcorrupt'
          - -avoid_negative_ts
          - make_zero
          - -strftime
          - '1'
          - -c:v
          - h264_vaapi
          - -c:a
          - aac

      hwaccel_args: preset-vaapi
      input_args: preset-rtsp-restream-low-latency
      retry_interval: 0
    record:
      enabled: true
      retain:
        days: 7
        mode: all
      alerts:
        retain:
          days: 90
        pre_capture: 10
        post_capture: 10
      detections:
        retain:
          days: 30
        pre_capture: 10
        post_capture: 10
    snapshots:
      enabled: true
      timestamp: false
      retain:
        default: 365
    audio:
      enabled: true
    motion:
      mask: 0.043,0.938,0.332,0.943,0.332,0.994,0.043,0.987
      threshold: 20
      contour_area: 10
      improve_contrast: false
    objects:
      filters:
        person: {}
version: 0.16-0
semantic_search:
  enabled: false
  reindex: false
  model_size: large

detectors:
  ov:
    type: openvino
    device: GPU

model:
  model_type: yolonas
  width: 640
  height: 640
  input_tensor: nchw
  input_pixel_format: bgr
  path: /config/yolo_nas_s.onnx
  labelmap_path: /labelmap/coco-80.txt
detect:
  enabled: true
face_recognition:
  enabled: false
  model_size: large
  detection_threshold: 0.4
  min_area: 200
lpr:
  enabled: false
classification:
  bird:
    enabled: false
tls:
  enabled: true
auth:
  enabled: false
notifications:
  enabled: true
  email: [email protected]
  cooldown: 10
proxy:
  default_role: admin
logger:
    logs:
        frigate.record.maintainer: debug

docker-compose file or Docker CLI command

services:
  frigate:
    image: ghcr.io/blakeblackshear/frigate:stable
    container_name: frigate
    shm_size: 256MB
    restart: unless-stopped
    devices:
      - /dev/dri:/dev/dri  # Intel GPU access for QSV decoding
    ports:
      - "5000:8971"        # Frigate Web UI
    volumes:
      - ./config:/config   # Config files
      - ./media:/media     # Recorded clips & snapshots
      - ../certs/pem:/etc/letsencrypt/live/frigate:ro
      - type: tmpfs
        target: /tmp/cache
        tmpfs:
          size: 1073741824
    environment:
      - LIBVA_DRIVER_NAME=iHD
    cap_add:
      - CAP_PERFMON

Relevant Frigate log output

2025-09-01 06:00:14.780538306  [2025-09-01 06:00:14] frigate.record.maintainer      DEBUG   : Copied /media/frigate/recordings/2025-09-01/06/cam_name/00.01.mp4 in 0.03493905067443848 seconds.
2025-09-01 06:00:24.773933440  [2025-09-01 06:00:24] frigate.record.maintainer      DEBUG   : Copied /media/frigate/recordings/2025-09-01/06/cam_name/00.11.mp4 in 0.029528141021728516 seconds.
2025-09-01 06:00:34.780749291  [2025-09-01 06:00:34] frigate.record.maintainer      DEBUG   : Copied /media/frigate/recordings/2025-09-01/06/cam_name/00.21.mp4 in 0.03063201904296875 seconds.
2025-09-01 06:00:44.779752064  [2025-09-01 06:00:44] frigate.record.maintainer      DEBUG   : Copied /media/frigate/recordings/2025-09-01/06/cam_name/00.31.mp4 in 0.03251314163208008 seconds.
2025-09-01 06:00:54.777190689  [2025-09-01 06:00:54] frigate.record.maintainer      DEBUG   : Copied /media/frigate/recordings/2025-09-01/06/cam_name/00.41.mp4 in 0.0318760871887207 seconds.
2025-09-01 06:01:04.760257884  [2025-09-01 06:01:04] frigate.record.maintainer      DEBUG   : Copied /media/frigate/recordings/2025-09-01/06/cam_name/00.51.mp4 in 0.022778987884521484 seconds.
2025-09-01 06:01:14.772363143  [2025-09-01 06:01:14] frigate.record.maintainer      DEBUG   : Copied /media/frigate/recordings/2025-09-01/06/cam_name/01.01.mp4 in 0.029525041580200195 seconds.
2025-09-01 06:01:24.779865312  [2025-09-01 06:01:24] frigate.record.maintainer      DEBUG   : Copied /media/frigate/recordings/2025-09-01/06/cam_name/01.11.mp4 in 0.030640840530395508 seconds.
2025-09-01 06:01:34.776361954  [2025-09-01 06:01:34] frigate.record.maintainer      DEBUG   : Copied /media/frigate/recordings/2025-09-01/06/cam_name/01.21.mp4 in 0.033660888671875 seconds.
2025-09-01 06:01:35.216773317  Exception in thread Thread-9:
2025-09-01 06:01:35.216775950  Traceback (most recent call last):
2025-09-01 06:01:35.216777173    File "/usr/lib/python3.11/threading.py", line 1038, in _bootstrap_inner
2025-09-01 06:01:35.216778008      self.run()
2025-09-01 06:01:35.216779023    File "/opt/frigate/frigate/video.py", line 244, in run
2025-09-01 06:01:35.216782003      while not self.stop_event.wait(self.sleeptime):
2025-09-01 06:01:35.216783866                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2025-09-01 06:01:35.216785454    File "/usr/lib/python3.11/multiprocessing/synchronize.py", line 349, in wait
2025-09-01 06:01:35.216811250      self._cond.wait(timeout)
2025-09-01 06:01:35.216812798    File "/usr/lib/python3.11/multiprocessing/synchronize.py", line 252, in wait
2025-09-01 06:01:35.216813975      self._sleeping_count.release()
2025-09-01 06:01:35.216815139  ValueError: semaphore or lock released too many times
2025-09-01 06:02:54.725964321  [2025-09-01 06:02:54] frigate.record.maintainer      WARNING : Too many unprocessed recording segments in cache for cam_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest...
2025-09-01 06:03:04.729809705  [2025-09-01 06:03:04] frigate.record.maintainer      WARNING : Too many unprocessed recording segments in cache for cam_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest...
2025-09-01 06:03:14.732399379  [2025-09-01 06:03:14] frigate.record.maintainer      WARNING : Too many unprocessed recording segments in cache for cam_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest...
2025-09-01 06:03:24.725923800  [2025-09-01 06:03:24] frigate.record.maintainer      WARNING : Too many unprocessed recording segments in cache for cam_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest...
2025-09-01 06:03:34.730981129  [2025-09-01 06:03:34] frigate.record.maintainer      WARNING : Too many unprocessed recording segments in cache for cam_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest...
2025-09-01 06:03:44.733319613  [2025-09-01 06:03:44] frigate.record.maintainer      WARNING : Too many unprocessed recording segments in cache for cam_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest...
2025-09-01 06:03:54.731079744  [2025-09-01 06:03:54] frigate.record.maintainer      WARNING : Too many unprocessed recording segments in cache for cam_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest...
2025-09-01 06:04:04.731180983  [2025-09-01 06:04:04] frigate.record.maintainer      WARNING : Too many unprocessed recording segments in cache for cam_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest...
2025-09-01 06:04:14.734161797  [2025-09-01 06:04:14] frigate.record.maintainer      WARNING : Too many unprocessed recording segments in cache for cam_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest...
2025-09-01 06:04:24.734367307  [2025-09-01 06:04:24] frigate.record.maintainer      WARNING : Too many unprocessed recording segments in cache for cam_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest...

Relevant go2rtc log output

2025-08-27 00:58:39.214016518  [INFO] Preparing new go2rtc config...
2025-08-27 00:58:39.606446176  [INFO] Starting go2rtc...
2025-08-27 00:58:39.676338164  00:58:39.676 INF go2rtc platform=linux/amd64 revision=fa580c5 version=1.9.9
2025-08-27 00:58:39.676353104  00:58:39.676 INF config path=/dev/shm/go2rtc.yaml
2025-08-27 00:58:39.676658432  00:58:39.676 INF [api] listen addr=:1984
2025-08-27 00:58:39.676661324  00:58:39.676 INF [rtsp] listen addr=:8554
2025-08-27 00:58:39.676957099  00:58:39.676 INF [webrtc] listen addr=:8555
2025-08-27 00:58:48.863835202  [INFO] Starting go2rtc healthcheck service...
2025-08-27 13:17:31.766325718  13:17:31.766 WRN github.com/AlexxIT/go2rtc/internal/streams/producer.go:170 > error="read tcp 172.20.0.2:60044->192.168.1.2:554: i/o timeout" url=rtsp://admin:[email protected]:554/h264_stream#timeout=2s
2025-08-27 13:17:43.950119613  13:17:43.950 WRN github.com/AlexxIT/go2rtc/internal/streams/producer.go:170 > error="read tcp 172.20.0.2:54584->192.168.1.2:554: i/o timeout" url=rtsp://admin:[email protected]:554/h264_stream#timeout=2s
2025-08-27 13:17:55.890987030  13:17:55.890 WRN github.com/AlexxIT/go2rtc/internal/streams/producer.go:170 > error="read tcp 172.20.0.2:60354->192.168.1.2:554: i/o timeout" url=rtsp://admin:[email protected]:554/h264_stream#timeout=2s

Operating system

Other

Install method

Docker Compose

Network connection

Wireless

Camera make and model

Ezviz CS-TY1-R101-1G2WF

Screenshots of the Frigate UI's System metrics pages

I already restarted Frigate to make recording continue, I didn't know I was going to be asked this.

Any other information that may be helpful

While my camera is cloud based, it offers an RTSP stream which is what made me consider using this project, since the Ezviz app is really bad at detection.

The attached go2rtc logged was downloaded a few days ago when the issue happened earlier, not when it just happened which is the one frigate's logs are from.

I know wireless is not recommended but I do not have the option currently. You will also notice my configuration has its own record parameters for ffmpeg. This is because I've been trying to fix an issue where clips attached video starts very late sometimes, like the event ended 10 seconds ago and the video is just 10 seconds of nothing after that. The custom flags helped a little bit, reduced the likelyhood of this issue and also reduced how often the camera's output printed time (which is very accurate btw, I checked) to somehow be ahead or behind the actual time frigate shows in history viewer.

Answered by NickM-27

Sep 1, 2025

I think your problem is the retry_interval of 0, it needs to be a number > 0

View full answer

2025-09-01T13:21:04Z

inkeep-ai-assistant[bot]
bot Sep 1, 2025

Based on the information you provided and the available documentation, the warning "Too many unprocessed recording segments in cache for cam_name. This likely indicates an issue with the detect stream, keeping the 6 most recent segments out of 7 and discarding the rest..." is triggered when Frigate's recording process cannot keep up with writing segments from RAM cache to disk(1). Normally, this is due to slow storage performance; however, your debug logs show that segments are being copied in under 100ms, so disk speed is not the issue.

The documentation suggests reviewing system resources such as CPU, RAM, disk throughput, and I/O bus utilization if you see this warning repeatedly(1). It also recommends monitoring with tools like docker stats, htop, iotop, iostat, and vmstat to check for any spikes or resource exhaustion at the time of failure(1).

The "Exception in thread" and "semaphore or lock released too many times" error is not explicitly documented, but it suggests a threading or synchronization problem—potentially a bug in Frigate's video processing thread. This is not a typical configuration or resource issue described in the docs.

Actions you can take based on the docs:

Continue to monitor your system using the recommended tools at the time just before the crash, to rule out resource exhaustion(1).
Ensure your Docker container's shm_size is large enough (you have 256MB set; if you have many cameras or high resolutions, try increasing this)(1).
If the problem persists and you continue to see the threading error, it may require developer attention, as this specific traceback is not addressed in the documentation.

Check out these potentially useful sources:
Troubleshooting Recordings,
Full Reference Config,
Advanced Options

Would you like a general answer based on my knowledge instead?

📚 Sources:

Troubleshooting Recordings

Have another question? Just tag @inkeep.

0 replies

NickM-27 · 2025-09-01T13:28:53Z

NickM-27
Sep 1, 2025
Collaborator Sponsor

I think your problem is the retry_interval of 0, it needs to be a number > 0

5 replies

BlueTree242 Sep 1, 2025
Author

I think your problem is the retry_interval of 0, it needs to be a number > 0

I originally added this in an attempt to reduce downtime as much as possible when camera automatically restarts/crashes/disconnects etc, is there a better way?

NickM-27 Sep 1, 2025
Collaborator Sponsor

The default values are optimal, this field is mainly there for people who need higher values for battery cameras

BlueTree242 Sep 1, 2025
Author

The default values are optimal, this field is mainly there for people who need higher values for battery cameras

So the default value will reconnect to my camera immediately when it's available? Or with a negligible delay?

BlueTree242 Sep 1, 2025
Author

I removed the option from my config immediately anyways.

BlueTree242 Sep 2, 2025
Author

Thank you for pointing this out very quickly, it fixed the issue as over 24 hours passed and frigate did not crash.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

[Support]: Frigate partly crashes after around 24 hours of uptime #19859

Uh oh!

{{title}}

Uh oh!

Replies: 2 comments 5 replies

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Select a reply

Uh oh!

Uh oh!

[Support]: Frigate partly crashes after around 24 hours of uptime #19859

Uh oh!

BlueTree242 Sep 1, 2025

Checklist

Describe the problem you are having

Steps to reproduce

Version

In which browser(s) are you experiencing the issue with?

Frigate config file

docker-compose file or Docker CLI command

Relevant Frigate log output

Relevant go2rtc log output

Operating system

Install method

Network connection

Camera make and model

Screenshots of the Frigate UI's System metrics pages

Any other information that may be helpful

Replies: 2 comments · 5 replies

Uh oh!

inkeep-ai-assistant[bot] bot Sep 1, 2025

Uh oh!

NickM-27 Sep 1, 2025 Collaborator Sponsor

Uh oh!

BlueTree242 Sep 1, 2025 Author

Uh oh!

NickM-27 Sep 1, 2025 Collaborator Sponsor

Uh oh!

BlueTree242 Sep 1, 2025 Author

Uh oh!

BlueTree242 Sep 1, 2025 Author

Uh oh!

BlueTree242 Sep 2, 2025 Author

BlueTree242
Sep 1, 2025

Replies: 2 comments 5 replies

inkeep-ai-assistant[bot]
bot Sep 1, 2025

NickM-27
Sep 1, 2025
Collaborator Sponsor

BlueTree242 Sep 1, 2025
Author

NickM-27 Sep 1, 2025
Collaborator Sponsor

BlueTree242 Sep 1, 2025
Author

BlueTree242 Sep 1, 2025
Author

BlueTree242 Sep 2, 2025
Author