CoZ Posted July 14, 2023 Share Posted July 14, 2023 Is there an “idiots guide” to setting this up? It took me about a month of troubleshooting and reading the docs on the older version and, I eventually got it up and running. I saw that an upgrade was available and there was no way I was going to attempt it after reading the docs as far as what was required to get the new version running. So, I left it alone. Unfortunately for me, I had to stop my Docker service to do some maintenance and the old version of Frigate, for whatever reason, would no longer launch. I made a backup of my config.yml file and attempted to install the new version to no avail. The docs aren't exactly idiot/user-friendly. There has got to be a better “how to” video/webpage out there that I can't find to assist in getting this new version back up and running? 1 Quote Link to comment
Pure Posted July 15, 2023 Share Posted July 15, 2023 (edited) Need some help, when trying to generate models using the tensorrt-models docker When trying to grab tensorrt_models.sh using the link in the template I get a "404: Not Found" error https://raw.githubusercontent.com/blakeblackshear/frigate/dev/docker/tensorrt_models.sh Using the link in the documentation https://docs.frigate.video/configuration/detectors https://github.com/blakeblackshear/frigate/raw/master/docker/tensorrt_models.sh open console and run the following chmod +x /mnt/user/appdata/trt-models/tensorrt_models.sh I can't execute or start the docker it gives me Execution error Bad parameter I get the following error in the logs text error warn system array login /opt/nvidia/nvidia_entrypoint.sh: /tensorrt_models.sh: /bin/bash^M: bad interpreter: No such file or directory /opt/nvidia/nvidia_entrypoint.sh: line 49: /tensorrt_models.sh: Success /opt/nvidia/nvidia_entrypoint.sh: /tensorrt_models.sh: /bin/bash^M: bad interpreter: No such file or directory /opt/nvidia/nvidia_entrypoint.sh: line 49: /tensorrt_models.sh: Success Edited July 15, 2023 by Pure Quote Link to comment
dopeytree Posted July 15, 2023 Share Posted July 15, 2023 (edited) Thanks so much for this container. Have it set up pretty well so far (ignore stationary objects & set threshold for object detection) Am using i9 11-900t 35w chip and a corel usb. As an example it averages around 40-60% cpu vs plex doing video thumbnails is using 130% cpu usage. Running full res detection on front of house and reduced on others. So that snapshots are better quality where needed. Once I have MQQT set up I might investigate using low res motion to take a still from the full res stream. Do any of you guys do this? mqtt: enabled: false host: 192.168.22.212 port: 1883 # Optional: topic prefix (default: shown below) # WARNING: must be unique if you are running multiple instances topic_prefix: frigate # Optional: client id (default: shown below) # WARNING: must be unique if you are running multiple instances client_id: frigate # Optional: user user: homeassistant # Optional: password # NOTE: Environment variables that begin with 'FRIGATE_' may be referenced in {}. # eg. password: '{FRIGATE_MQTT_PASSWORD}' password: x # Optional: interval in seconds for publishing stats (default: shown below) stats_interval: 60 go2rtc: streams: barn_inside: - rtsp://pSrUq7TZ:[email protected]:554/live/ch0 house_front_door: - rtsp://EWLHSHKY:[email protected]:554/live/ch0 ffmpeg: output_args: record: preset-record-generic-audio-aac hwaccel_args: - -hwaccel - qsv - -qsv_device - /dev/dri/renderD128 - -hwaccel_output_format - yuv420p detectors: coral: type: edgetpu device: usb cameras: barn_inside: snapshots: enabled: true ffmpeg: output_args: record: preset-record-generic-audio-aac hwaccel_args: preset-intel-qsv-h264 inputs: - path: rtsp://pSrUq7TZ:[email protected]:554/live/ch1 roles: - detect - path: rtsp://192.168.22.2:8554/barn_inside?mp4 roles: - record detect: width: 768 height: 432 fps: 5 stationary: interval: 0 threshold: 70 objects: track: - person - cat - bird - cell phone - banana - wine glass - backpack record: enabled: True retain: days: 7 mode: motion events: retain: default: 14 mode: active_objects house_front_door: snapshots: enabled: true ffmpeg: output_args: record: preset-record-generic-audio-aac hwaccel_args: preset-intel-qsv-h264 inputs: - path: rtsp://EWLHSHKY:[email protected]:554/live/ch0 roles: - detect - path: rtsp://192.168.22.2:8554/house_front_door?mp4 roles: - record detect: width: 1920 height: 1080 fps: 5 stationary: interval: 0 threshold: 70 motion: mask: - 956,720,1280,720,1280,409,1174,363,1036,419,956,527 - 891,200,985,229,1140,204,1189,117,1137,59,916,46,782,122 - 1280,0,1280,37,874,40,872,0 - 690,146,690,212,513,216,514,150 - 268,399,333,497,523,480,515,359,403,275 - 412,720,637,699,474,604,363,518,275,533,223,570,311,656 objects: track: - person - car - dog - cat - bird - bicycle - motorcycle - cell phone - backpack - horse filters: car: threshold: 0.7 min_area: 1000 max_area: 100000 person: threshold: 0.7 min_area: 1000 max_area: 100000 record: enabled: True retain: days: 7 mode: motion events: retain: default: 14 mode: active_objects Edited July 15, 2023 by dopeytree Quote Link to comment
yayitazale Posted July 16, 2023 Author Share Posted July 16, 2023 On 7/14/2023 at 3:38 PM, MothyTim said: Hi, I had a frustrating day yesterday trying to upgrade to v0.12 gave up in the end and rolled back to v0.11.1! I went through the release notes and changed the things I thought and while the docker image was running I had no acces the the gui and the logs wouldn't even open! It was like the log window would crash as soon as it opened! So without a log to see what the errors were I was stuck! Please if someone has time to look at my current config file and point out the bits I need to change, then I'll have another go at it! Cheers Tim config.yml 9.03 kB · 4 downloads Please you should read the docs, you should change a lot of thing so I can't tell you every single thing if you don't at least read the documentation that is pretty complete. Quote Link to comment
yayitazale Posted July 16, 2023 Author Share Posted July 16, 2023 On 7/14/2023 at 9:13 PM, CoZ said: Is there an “idiots guide” to setting this up? It took me about a month of troubleshooting and reading the docs on the older version and, I eventually got it up and running. I saw that an upgrade was available and there was no way I was going to attempt it after reading the docs as far as what was required to get the new version running. So, I left it alone. Unfortunately for me, I had to stop my Docker service to do some maintenance and the old version of Frigate, for whatever reason, would no longer launch. I made a backup of my config.yml file and attempted to install the new version to no avail. The docs aren't exactly idiot/user-friendly. There has got to be a better “how to” video/webpage out there that I can't find to assist in getting this new version back up and running? Blame/ask for help in the correct place if you have doubts about the container config itself: https://github.com/blakeblackshear/frigate/discussions Quote Link to comment
yayitazale Posted July 16, 2023 Author Share Posted July 16, 2023 On 7/15/2023 at 12:11 PM, Pure said: Need some help, when trying to generate models using the tensorrt-models docker When trying to grab tensorrt_models.sh using the link in the template I get a "404: Not Found" error https://raw.githubusercontent.com/blakeblackshear/frigate/dev/docker/tensorrt_models.sh Using the link in the documentation https://docs.frigate.video/configuration/detectors https://github.com/blakeblackshear/frigate/raw/master/docker/tensorrt_models.sh open console and run the following chmod +x /mnt/user/appdata/trt-models/tensorrt_models.sh I can't execute or start the docker it gives me Execution error Bad parameter I get the following error in the logs text error warn system array login /opt/nvidia/nvidia_entrypoint.sh: /tensorrt_models.sh: /bin/bash^M: bad interpreter: No such file or directory /opt/nvidia/nvidia_entrypoint.sh: line 49: /tensorrt_models.sh: Success /opt/nvidia/nvidia_entrypoint.sh: /tensorrt_models.sh: /bin/bash^M: bad interpreter: No such file or directory /opt/nvidia/nvidia_entrypoint.sh: line 49: /tensorrt_models.sh: Success Does "chmod +x /mnt/user/appdata/trt-models/tensorrt_models.sh" give you some output? Do you have the /mnt/user/appdata/trt-models folder mapped already to frigate and frigate running? Quote Link to comment
yayitazale Posted July 16, 2023 Author Share Posted July 16, 2023 21 hours ago, dopeytree said: Thanks so much for this container. Have it set up pretty well so far (ignore stationary objects & set threshold for object detection) Am using i9 11-900t 35w chip and a corel usb. As an example it averages around 40-60% cpu vs plex doing video thumbnails is using 130% cpu usage. Running full res detection on front of house and reduced on others. So that snapshots are better quality where needed. Once I have MQQT set up I might investigate using low res motion to take a still from the full res stream. Do any of you guys do this? mqtt: enabled: false host: 192.168.22.212 port: 1883 # Optional: topic prefix (default: shown below) # WARNING: must be unique if you are running multiple instances topic_prefix: frigate # Optional: client id (default: shown below) # WARNING: must be unique if you are running multiple instances client_id: frigate # Optional: user user: homeassistant # Optional: password # NOTE: Environment variables that begin with 'FRIGATE_' may be referenced in {}. # eg. password: '{FRIGATE_MQTT_PASSWORD}' password: x # Optional: interval in seconds for publishing stats (default: shown below) stats_interval: 60 go2rtc: streams: barn_inside: - rtsp://pSrUq7TZ:[email protected]:554/live/ch0 house_front_door: - rtsp://EWLHSHKY:[email protected]:554/live/ch0 ffmpeg: output_args: record: preset-record-generic-audio-aac hwaccel_args: - -hwaccel - qsv - -qsv_device - /dev/dri/renderD128 - -hwaccel_output_format - yuv420p detectors: coral: type: edgetpu device: usb cameras: barn_inside: snapshots: enabled: true ffmpeg: output_args: record: preset-record-generic-audio-aac hwaccel_args: preset-intel-qsv-h264 inputs: - path: rtsp://pSrUq7TZ:[email protected]:554/live/ch1 roles: - detect - path: rtsp://192.168.22.2:8554/barn_inside?mp4 roles: - record detect: width: 768 height: 432 fps: 5 stationary: interval: 0 threshold: 70 objects: track: - person - cat - bird - cell phone - banana - wine glass - backpack record: enabled: True retain: days: 7 mode: motion events: retain: default: 14 mode: active_objects house_front_door: snapshots: enabled: true ffmpeg: output_args: record: preset-record-generic-audio-aac hwaccel_args: preset-intel-qsv-h264 inputs: - path: rtsp://EWLHSHKY:[email protected]:554/live/ch0 roles: - detect - path: rtsp://192.168.22.2:8554/house_front_door?mp4 roles: - record detect: width: 1920 height: 1080 fps: 5 stationary: interval: 0 threshold: 70 motion: mask: - 956,720,1280,720,1280,409,1174,363,1036,419,956,527 - 891,200,985,229,1140,204,1189,117,1137,59,916,46,782,122 - 1280,0,1280,37,874,40,872,0 - 690,146,690,212,513,216,514,150 - 268,399,333,497,523,480,515,359,403,275 - 412,720,637,699,474,604,363,518,275,533,223,570,311,656 objects: track: - person - car - dog - cat - bird - bicycle - motorcycle - cell phone - backpack - horse filters: car: threshold: 0.7 min_area: 1000 max_area: 100000 person: threshold: 0.7 min_area: 1000 max_area: 100000 record: enabled: True retain: days: 7 mode: motion events: retain: default: 14 mode: active_objects I think you can't but you can open a discussion here: https://github.com/blakeblackshear/frigate/discussions Quote Link to comment
MothyTim Posted July 16, 2023 Share Posted July 16, 2023 3 hours ago, yayitazale said: Please you should read the docs, you should change a lot of thing so I can't tell you every single thing if you don't at least read the documentation that is pretty complete. Ok, did that for hours! I'll stick with v11 as it works! Whilst I appreciate all the work people put into these programs, I find the breaking changes very frustrating, especially when there are so many to work out at once! So I'll move to Unifi Protect as soon as I can afford the cameras, as it has a GUI. Thanks again for your efforts bringing this to Unraid! 1 Quote Link to comment
CoZ Posted July 17, 2023 Share Posted July 17, 2023 (edited) 9 hours ago, yayitazale said: Blame/ask for help in the correct place if you have doubts about the container config itself: https://github.com/blakeblackshear/frigate/discussions So, I've been messing with it because I am too stubborn to call it quits. I seem to have it up and running. However, picking the Nvidia Branch when selecting which version to install from CA and then trying to enter into the config.yml: hwaccel_args: preset-nvidia-h264 Which, causes an error about FFMPEG crashing. If I remove the hwaccel_args: preset-nvidia-h264 and restart frigate docker, I can see the stream. I did enter my GPUID and clicked advancded tab and entered the --runtime=nvidia at the end of all the other text that was already pre-populated. --shm-size=256mb --mount type=tmpfs,target=/tmp/cache,tmpfs-size=100000000 --restart unless-stopped --runtime=nvidia Running the command nvidia-smi without the hwaccel_args tells me that there are no running processes. I'm using a Geforce GTX 1050 on a Reolink RLC-410W You DID tell me to ask for help. So, this is pretty much your fault now Edited July 17, 2023 by CoZ left out some info Quote Link to comment
yayitazale Posted July 17, 2023 Author Share Posted July 17, 2023 10 hours ago, MothyTim said: Ok, did that for hours! I'll stick with v11 as it works! Whilst I appreciate all the work people put into these programs, I find the breaking changes very frustrating, especially when there are so many to work out at once! So I'll move to Unifi Protect as soon as I can afford the cameras, as it has a GUI. Thanks again for your efforts bringing this to Unraid! Watching your config file, thing you should change: RTMP is deprecated, so you should start using the new restream, here the info: https://docs.frigate.video/configuration/restream Also you should change the camera role definition to the new scheme using the restream: https://docs.frigate.video/configuration/cameras Maybe you need to change a little bit also the events definition on cameras (not sure, maybe not): https://docs.frigate.video/configuration/record#what-do-the-different-retain-modes-mean You can change the FFMPEG args to presets: https://docs.frigate.video/configuration/hardware_acceleration#intel-based-cpus I suggest you to star from scratch with a new config file with a minimal configuration as it's recommended here: https://docs.frigate.video/configuration/ Then start adding or changing specific things one by one to avoid multiple errors at the same time that are harder to tackle. 1 Quote Link to comment
yayitazale Posted July 17, 2023 Author Share Posted July 17, 2023 (edited) 5 hours ago, CoZ said: So, I've been messing with it because I am too stubborn to call it quits. I seem to have it up and running. However, picking the Nvidia Branch when selecting which version to install from CA and then trying to enter into the config.yml: hwaccel_args: preset-nvidia-h264 Which, causes an error about FFMPEG crashing. If I remove the hwaccel_args: preset-nvidia-h264 and restart frigate docker, I can see the stream. I did enter my GPUID and clicked advancded tab and entered the --runtime=nvidia at the end of all the other text that was already pre-populated. --shm-size=256mb --mount type=tmpfs,target=/tmp/cache,tmpfs-size=100000000 --restart unless-stopped --runtime=nvidia Running the command nvidia-smi without the hwaccel_args tells me that there are no running processes. I'm using a Geforce GTX 1050 on a Reolink RLC-410W You DID tell me to ask for help. So, this is pretty much your fault now Are you going to use Nvidia GPU as a detector? If it is not the case, you should use the default branch. You should add "--runtime=nvidia" as first extra parameter, not last. Can you post the error of the FFMPEG? Where are you running the nvidia-smi command? You have to run it in the unraid OS console, you will not see anything inside the container. Edited July 17, 2023 by yayitazale Quote Link to comment
Pure Posted July 17, 2023 Share Posted July 17, 2023 17 hours ago, yayitazale said: Does "chmod +x /mnt/user/appdata/trt-models/tensorrt_models.sh" give you some output? Do you have the /mnt/user/appdata/trt-models folder mapped already to frigate and frigate running? Chmod +c gave no output Had the folder mapped and frigate running Quote Link to comment
yayitazale Posted July 17, 2023 Author Share Posted July 17, 2023 2 hours ago, Pure said: Chmod +c gave no output Had the folder mapped and frigate running Can you try with frigate stopped? Quote Link to comment
CoZ Posted July 17, 2023 Share Posted July 17, 2023 (edited) 16 hours ago, yayitazale said: Are you going to use Nvidia GPU as a detector? If it is not the case, you should use the default branch. You should add "--runtime=nvidia" as first extra parameter, not last. Can you post the error of the FFMPEG? Where are you running the nvidia-smi command? You have to run it in the unraid OS console, you will not see anything inside the container. 1. Eventually – I guess? I have one camera and I only use it to detect when my family of wild raccoons show up at night so I can go out and feed them and hang with them. Please, don't judge. lol I know I have to download another docker image Tensor (something or other) to utilize it. I'm just concentrating on getting this back up and running. 2. Done. They didn't specify if it should be before or after. Again, the DOCS leave a lot to be desired for amateurs such as I. 3. 2023-07-17 14:28:52.680503315 [2023-07-17 14:28:52] frigate.video ERROR : Rear_Deck: Unable to read frames from ffmpeg process. 2023-07-17 14:28:52.680790112 [2023-07-17 14:28:52] frigate.video ERROR : Rear_Deck: ffmpeg process is not running. exiting capture thread... 2023-07-17 14:28:55.961193412 [2023-07-17 14:28:55] watchdog.Rear_Deck ERROR : Ffmpeg process crashed unexpectedly for Rear_Deck. 2023-07-17 14:28:55.961446594 [2023-07-17 14:28:55] watchdog.Rear_Deck ERROR : The following ffmpeg logs include the last 100 lines prior to exit. 2023-07-17 14:28:55.961725123 [2023-07-17 14:28:55] ffmpeg.Rear_Deck.detect ERROR : [segment @ 0x55dfd77aff40] Timestamps are unset in a packet for stream 0. This is deprecated and will stop working in the future. Fix your code to set the timestamps properly 2023-07-17 14:28:55.961920823 [2023-07-17 14:28:55] ffmpeg.Rear_Deck.detect ERROR : [rtsp @ 0x55dfd7793200] CSeq 6 expected, 0 received. 2023-07-17 14:28:55.962207303 [2023-07-17 14:28:55] ffmpeg.Rear_Deck.detect ERROR : Last message repeated 4 times 2023-07-17 14:28:55.962494096 [2023-07-17 14:28:55] ffmpeg.Rear_Deck.detect ERROR : rtsp://*:*@192.168.1.18:554/h264Preview_01_main: Connection timed out 2023-07-17 14:28:55.962704420 [2023-07-17 14:28:55] ffmpeg.Rear_Deck.detect ERROR : Last message repeated 1 times 4. I know enough to run it from the CLI in UInraid (Plex and Jellyfin and Unimanic user). Otherwise, that piece of information would have been useful. New information: After switching the runtime=nvidia to the front of the extra parameter line and placing back in the hwaccel_args: preset-nvidia-h264 back into my config.yml. Even though it explodes and I get all those FFMPEG error messages I can now actually VIEW the camera stream and typing in nvidia-smi into the CLI shows that a FFMPEG running process is ... running. So, no idea why or how? I leave that up to you good Sir. Update 19:46HRS: Tried messing with the Tensor docker. It filled my docker memory usage to 90% once it extracted almost 4GB worth of files. Almost crashed my entire system. Not good. Edited July 17, 2023 by CoZ Quote Link to comment
yayitazale Posted July 18, 2023 Author Share Posted July 18, 2023 12 hours ago, CoZ said: 1. Eventually – I guess? I have one camera and I only use it to detect when my family of wild raccoons show up at night so I can go out and feed them and hang with them. Please, don't judge. lol I know I have to download another docker image Tensor (something or other) to utilize it. I'm just concentrating on getting this back up and running. 2. Done. They didn't specify if it should be before or after. Again, the DOCS leave a lot to be desired for amateurs such as I. 3. 2023-07-17 14:28:52.680503315 [2023-07-17 14:28:52] frigate.video ERROR : Rear_Deck: Unable to read frames from ffmpeg process. 2023-07-17 14:28:52.680790112 [2023-07-17 14:28:52] frigate.video ERROR : Rear_Deck: ffmpeg process is not running. exiting capture thread... 2023-07-17 14:28:55.961193412 [2023-07-17 14:28:55] watchdog.Rear_Deck ERROR : Ffmpeg process crashed unexpectedly for Rear_Deck. 2023-07-17 14:28:55.961446594 [2023-07-17 14:28:55] watchdog.Rear_Deck ERROR : The following ffmpeg logs include the last 100 lines prior to exit. 2023-07-17 14:28:55.961725123 [2023-07-17 14:28:55] ffmpeg.Rear_Deck.detect ERROR : [segment @ 0x55dfd77aff40] Timestamps are unset in a packet for stream 0. This is deprecated and will stop working in the future. Fix your code to set the timestamps properly 2023-07-17 14:28:55.961920823 [2023-07-17 14:28:55] ffmpeg.Rear_Deck.detect ERROR : [rtsp @ 0x55dfd7793200] CSeq 6 expected, 0 received. 2023-07-17 14:28:55.962207303 [2023-07-17 14:28:55] ffmpeg.Rear_Deck.detect ERROR : Last message repeated 4 times 2023-07-17 14:28:55.962494096 [2023-07-17 14:28:55] ffmpeg.Rear_Deck.detect ERROR : rtsp://*:*@192.168.1.18:554/h264Preview_01_main: Connection timed out 2023-07-17 14:28:55.962704420 [2023-07-17 14:28:55] ffmpeg.Rear_Deck.detect ERROR : Last message repeated 1 times 4. I know enough to run it from the CLI in UInraid (Plex and Jellyfin and Unimanic user). Otherwise, that piece of information would have been useful. New information: After switching the runtime=nvidia to the front of the extra parameter line and placing back in the hwaccel_args: preset-nvidia-h264 back into my config.yml. Even though it explodes and I get all those FFMPEG error messages I can now actually VIEW the camera stream and typing in nvidia-smi into the CLI shows that a FFMPEG running process is ... running. So, no idea why or how? I leave that up to you good Sir. Update 19:46HRS: Tried messing with the Tensor docker. It filled my docker memory usage to 90% once it extracted almost 4GB worth of files. Almost crashed my entire system. Not good. 1. I wasn't judging, only pointing out something that some people doesn't understand or even read. 2. This is something basic of docker, not something exclusive from frigate. In any case, if you have complains about the docs you are welcome to send a PR with the suggested changes to the template: https://gitlab.com/yayitazale/unraid-templates/-/blob/main/yayitazale/frigate.xml or to the docs of the project: https://github.com/blakeblackshear/frigate/tree/dev/docs 3. Can you attach a screenshot of the template config you are using and the config file? You can expand the docker image from unraid if you are out of space, this is something that I can't expect from every user. You are the responsible of your system not me. If you don't like this just buy any commercial software. Quote Link to comment
CoZ Posted July 18, 2023 Share Posted July 18, 2023 6 hours ago, yayitazale said: 1. I wasn't judging, only pointing out something that some people doesn't understand or even read. 2. This is something basic of docker, not something exclusive from frigate. In any case, if you have complains about the docs you are welcome to send a PR with the suggested changes to the template: https://gitlab.com/yayitazale/unraid-templates/-/blob/main/yayitazale/frigate.xml or to the docs of the project: https://github.com/blakeblackshear/frigate/tree/dev/docs 3. Can you attach a screenshot of the template config you are using and the config file? You can expand the docker image from unraid if you are out of space, this is something that I can't expect from every user. You are the responsible of your system not me. If you don't like this just buy any commercial software. I've expanded my docker.img a few times now. I am also partially to blame. Pulling down your container the first time and then going back in again and changing the repo (after actually digesting your instructions) to the newest Nvidia one and then pulling all that data again became far too much. I never did get it working. I saved your script and tried to place it in to the appdata/trt-models/tensorrt_models.sh folder. It would not let me because the folder is the same as the filename. So then I tried to change the name of the script and then change the whole chmod stuff to the new name of the script - still didn't work. I know you wren't judging. I was trying to bring some levity into the discussion. You want me to post my config.yml or, you want a screenshot of how I have your container setup with all the paths and etc...etc..? Remember, talk to me as if you were talking to a small child trying to explain how the universe was created. (again, levity) Quote Link to comment
yayitazale Posted July 18, 2023 Author Share Posted July 18, 2023 5 minutes ago, CoZ said: I've expanded my docker.img a few times now. I am also partially to blame. Pulling down your container the first time and then going back in again and changing the repo (after actually digesting your instructions) to the newest Nvidia one and then pulling all that data again became far too much. I never did get it working. I saved your script and tried to place it in to the appdata/trt-models/tensorrt_models.sh folder. It would not let me because the folder is the same as the filename. So then I tried to change the name of the script and then change the whole chmod stuff to the new name of the script - still didn't work. I know you wren't judging. I was trying to bring some levity into the discussion. You want me to post my config.yml or, you want a screenshot of how I have your container setup with all the paths and etc...etc..? Remember, talk to me as if you were talking to a small child trying to explain how the universe was created. (again, levity) First of all, to your info, I only maintain the template for unraid. I'm not a developer of frigate so the script is not mine, and I only translated the deploying instruction from pure docker of the frigate docs to a unraid app template. I would like to have both, the screenshot and the file, to see if there is something to fix in the way you are launching the container or in the config file. 1 Quote Link to comment
CoZ Posted July 18, 2023 Share Posted July 18, 2023 As requested. I am sure you are going to have a field day with this one and just rip apart what I've done. I'm ready. Go for it. config.yml Quote Link to comment
yayitazale Posted July 19, 2023 Author Share Posted July 19, 2023 14 hours ago, CoZ said: As requested. I am sure you are going to have a field day with this one and just rip apart what I've done. I'm ready. Go for it. config.yml 863 B · 2 downloads I think you should add a input arg: input_args: preset-rtsp-generic And for output, you should use a preset too: output_args: preset-record-generic In any case, http video streams seem to be the most reliable for Reolink, there you have some info about how to setup the got2rtc restreaming: https://docs.frigate.video/configuration/camera_specific#reolink-cameras For more doubts about your config, please see https://github.com/blakeblackshear/frigate/issues/3235#issuecomment-1135876973 and you can ask there. Quote Link to comment
CoZ Posted July 19, 2023 Share Posted July 19, 2023 3 hours ago, yayitazale said: In any case, http video streams seem to be the most reliable for Reolink, there you have some info about how to setup the got2rtc restreaming: https://docs.frigate.video/configuration/camera_specific#reolink-cameras That was part of the initial docker pull - the got2rtc streaming you're speaking of? I deleted all those entries on the docker because (A) had no clue what it was for (B) it told me to remove them if they weren't going to be used. Putting in the output_args caused Frigate to crash with the following error: 2023-07-19 09:31:52.056562168 pydantic.error_wrappers.ValidationError: 1 validation error for FrigateConfig 2023-07-19 09:31:52.056566211 cameras -> Rear_Deck -> ffmpeg -> output_args -> output_args 2023-07-19 09:31:52.056569915 extra fields not permitted (type=value_error.extra) Quote Link to comment
yayitazale Posted July 20, 2023 Author Share Posted July 20, 2023 23 hours ago, CoZ said: That was part of the initial docker pull - the got2rtc streaming you're speaking of? I deleted all those entries on the docker because (A) had no clue what it was for (B) it told me to remove them if they weren't going to be used. Putting in the output_args caused Frigate to crash with the following error: 2023-07-19 09:31:52.056562168 pydantic.error_wrappers.ValidationError: 1 validation error for FrigateConfig 2023-07-19 09:31:52.056566211 cameras -> Rear_Deck -> ffmpeg -> output_args -> output_args 2023-07-19 09:31:52.056569915 extra fields not permitted (type=value_error.extra) No, I'm talking about the config file. You didn't read the docs properly: https://docs.frigate.video/configuration/restream The log points clearly that you are not nesting the config file correctly. It should be cameras -> Rear_Deck -> ffmpeg -> output_args: preset-record-generic Quote Link to comment
CoZ Posted July 22, 2023 Share Posted July 22, 2023 (edited) On 7/20/2023 at 8:55 AM, yayitazale said: No, I'm talking about the config file. You didn't read the docs properly: https://docs.frigate.video/configuration/restream Sorry for the delay in replying. Some severe weather rolled through, and I was stuck at various airports forever. My friend, if I could UNDERSTAND the Docs and read them properly, I wouldn't be here asking for help. That is my main complaint with this. It's by no means 'user/idiot friendly'. There are too many variables listed in those docs. Hence, why I was looking for a “If you have this model of camera, your config file needs to look like this” type of website earlier. You have my gratitude for helping me get this far. You also have my sympathies for constantly trying to help others on here. Home Assistant has better docs / wiki and support forum. I didn't really have any issues setting all that up and the associated containers (NodeRed, MQTT, Zwave, etc) since it was easy to follow along. For some reason, I can't wrap my head around Frigate. Edited July 22, 2023 by CoZ Quote Link to comment
MothyTim Posted July 24, 2023 Share Posted July 24, 2023 On 7/17/2023 at 7:50 AM, yayitazale said: Watching your config file, thing you should change: RTMP is deprecated, so you should start using the new restream, here the info: https://docs.frigate.video/configuration/restream Also you should change the camera role definition to the new scheme using the restream: https://docs.frigate.video/configuration/cameras Maybe you need to change a little bit also the events definition on cameras (not sure, maybe not): https://docs.frigate.video/configuration/record#what-do-the-different-retain-modes-mean You can change the FFMPEG args to presets: https://docs.frigate.video/configuration/hardware_acceleration#intel-based-cpus I suggest you to star from scratch with a new config file with a minimal configuration as it's recommended here: https://docs.frigate.video/configuration/ Then start adding or changing specific things one by one to avoid multiple errors at the same time that are harder to tackle. Thanks for the tips! I put my calm head on and had another go yesterday while watching the Grand Prix! All worked straight away! I copied the newer Reolink config over and switched off RTMP and left everthing else mostly the same! One question! Do you think its better to change the FFMEG settings to one of the presets? Cheers, Timconfig.yml Quote Link to comment
yayitazale Posted July 26, 2023 Author Share Posted July 26, 2023 On 7/24/2023 at 1:02 PM, MothyTim said: Thanks for the tips! I put my calm head on and had another go yesterday while watching the Grand Prix! All worked straight away! I copied the newer Reolink config over and switched off RTMP and left everthing else mostly the same! One question! Do you think its better to change the FFMEG settings to one of the presets? Cheers, Timconfig.yml Yes, presets will simplify your life. In your case you should delete the ffmpeg part and add only: ffmpeg: hwaccel_args: preset-intel-qsv-h264 Or this, if qsv doesn't work: ffmpeg: hwaccel_args: preset-vaapi In my case, I also add this to the ffmpeg part to work with webrtc in low latency and to record with the transcoded audio: input_args: preset-rtsp-restream-low-latency output_args: record: preset-record-generic-audio-copy And this in the go2rtc part in order to be able to connect from Home Assistant: webrtc: candidates: - 192.168.X.XX:8555 - stun:8555 I see that you should also simplify the definition of cameras. You defined the substream that you are not using, and also, you are transcoding to opus and then to acc. I think you shoud only transocde once and keep that, so you should have something like this (chnage the opus to aac if needed): go2rtc: streams: garden_camera: - http://10.19.64.220/flv?port=1935&app=bcs&stream=channel0_main.bcs&user=admin&password={FRIGATE_RTSP_PASSWORD} - "ffmpeg:garden_camera#audio=opus" cameras: garden: ffmpeg: inputs: - path: rtsp://127.0.0.1:8554/garden_camera input_args: preset-rtsp-restream roles: - record - detect For this last part, you should have defined the ffpmeg > output_args > records as I told you to preset-record-generic-audio-copy to record with the opus/aac (whatever you keep with) audio. In any case, you can also maintain the flv stream to detection and only restream for recording, you have the specific example for reolink config to do that here: https://docs.frigate.video/configuration/ffmpeg_presets#input-args-presets 1 Quote Link to comment
MothyTim Posted July 26, 2023 Share Posted July 26, 2023 2 hours ago, yayitazale said: Yes, presets will simplify your life. In your case you should delete the ffmpeg part and add only: ffmpeg: hwaccel_args: preset-intel-qsv-h264 Or this, if qsv doesn't work: ffmpeg: hwaccel_args: preset-vaapi In my case, I also add this to the ffmpeg part to work with webrtc in low latency and to record with the transcoded audio: input_args: preset-rtsp-restream-low-latency output_args: record: preset-record-generic-audio-copy And this in the go2rtc part in order to be able to connect from Home Assistant: webrtc: candidates: - 192.168.X.XX:8555 - stun:8555 I see that you should also simplify the definition of cameras. You defined the substream that you are not using, and also, you are transcoding to opus and then to acc. I think you shoud only transocde once and keep that, so you should have something like this (chnage the opus to aac if needed): go2rtc: streams: garden_camera: - http://10.19.64.220/flv?port=1935&app=bcs&stream=channel0_main.bcs&user=admin&password={FRIGATE_RTSP_PASSWORD} - "ffmpeg:garden_camera#audio=opus" cameras: garden: ffmpeg: inputs: - path: rtsp://127.0.0.1:8554/garden_camera input_args: preset-rtsp-restream roles: - record - detect For this last part, you should have defined the ffpmeg > output_args > records as I told you to preset-record-generic-audio-copy to record with the opus/aac (whatever you keep with) audio. In any case, you can also maintain the flv stream to detection and only restream for recording, you have the specific example for reolink config to do that here: https://docs.frigate.video/configuration/ffmpeg_presets#input-args-presets Hi, thanks for all that. I've changed the things that you've suggested, but unfortunatly it doesn't work? Here is the log and new config. text error warn system array login Traceback (most recent call last): File "/usr/local/go2rtc/create_config.py", line 27, in <module> config: dict[str, any] = yaml.safe_load(raw_config) File "/usr/local/lib/python3.9/dist-packages/yaml/__init__.py", line 125, in safe_load return load(stream, SafeLoader) File "/usr/local/lib/python3.9/dist-packages/yaml/__init__.py", line 81, in load return loader.get_single_data() File "/usr/local/lib/python3.9/dist-packages/yaml/constructor.py", line 49, in get_single_data node = self.get_single_node() File "/usr/local/lib/python3.9/dist-packages/yaml/composer.py", line 36, in get_single_node document = self.compose_document() File "/usr/local/lib/python3.9/dist-packages/yaml/composer.py", line 55, in compose_document node = self.compose_node(None, None) File "/usr/local/lib/python3.9/dist-packages/yaml/composer.py", line 84, in compose_node node = self.compose_mapping_node(anchor) File "/usr/local/lib/python3.9/dist-packages/yaml/composer.py", line 133, in compose_mapping_node item_value = self.compose_node(node, item_key) File "/usr/local/lib/python3.9/dist-packages/yaml/composer.py", line 84, in compose_node node = self.compose_mapping_node(anchor) File "/usr/local/lib/python3.9/dist-packages/yaml/composer.py", line 127, in compose_mapping_node while not self.check_event(MappingEndEvent): File "/usr/local/lib/python3.9/dist-packages/yaml/parser.py", line 98, in check_event self.current_event = self.state() File "/usr/local/lib/python3.9/dist-packages/yaml/parser.py", line 428, in parse_block_mapping_key if self.check_token(KeyToken): File "/usr/local/lib/python3.9/dist-packages/yaml/scanner.py", line 116, in check_token self.fetch_more_tokens() File "/usr/local/lib/python3.9/dist-packages/yaml/scanner.py", line 223, in fetch_more_tokens return self.fetch_value() File "/usr/local/lib/python3.9/dist-packages/yaml/scanner.py", line 577, in fetch_value raise ScannerError(None, None, yaml.scanner.ScannerError: mapping values are not allowed here in "<unicode string>", line 180, column 15: input_args: preset-rtsp-restream-low-latency ^ s6-rc: info: service legacy-services: stopping s6-rc: info: service legacy-services successfully stopped s6-rc: info: service nginx: stopping s6-rc: info: service go2rtc-healthcheck: stopping s6-rc: info: service go2rtc-healthcheck successfully stopped s6-rc: info: service nginx successfully stopped s6-rc: info: service nginx-log: stopping s6-rc: info: service frigate: stopping s6-rc: info: service frigate successfully stopped s6-rc: info: service go2rtc: stopping s6-rc: info: service frigate-log: stopping s6-rc: info: service nginx-log successfully stopped s6-rc: info: service go2rtc successfully stopped s6-rc: info: service go2rtc-log: stopping s6-rc: info: service frigate-log successfully stopped s6-rc: info: service go2rtc-log successfully stopped s6-rc: info: service log-prepare: stopping s6-rc: info: service s6rc-fdholder: stopping s6-rc: info: service log-prepare successfully stopped s6-rc: info: service legacy-cont-init: stopping s6-rc: info: service s6rc-fdholder successfully stopped s6-rc: info: service legacy-cont-init successfully stopped s6-rc: info: service fix-attrs: stopping s6-rc: info: service fix-attrs successfully stopped s6-rc: info: service s6rc-oneshot-runner: stopping s6-rc: info: service s6rc-oneshot-runner successfully stopped 2023-07-26 12:15:22.225758077 File "/usr/local/lib/python3.9/dist-packages/yaml/composer.py", line 84, in compose_node 2023-07-26 12:15:22.225759078 node = self.compose_mapping_node(anchor) 2023-07-26 12:15:22.225760116 File "/usr/local/lib/python3.9/dist-packages/yaml/composer.py", line 127, in compose_mapping_node 2023-07-26 12:15:22.225761077 while not self.check_event(MappingEndEvent): 2023-07-26 12:15:22.225762094 File "/usr/local/lib/python3.9/dist-packages/yaml/parser.py", line 98, in check_event 2023-07-26 12:15:22.225762944 self.current_event = self.state() 2023-07-26 12:15:22.225763977 File "/usr/local/lib/python3.9/dist-packages/yaml/parser.py", line 428, in parse_block_mapping_key 2023-07-26 12:15:22.225781265 if self.check_token(KeyToken): 2023-07-26 12:15:22.225782421 File "/usr/local/lib/python3.9/dist-packages/yaml/scanner.py", line 116, in check_token 2023-07-26 12:15:22.225783244 self.fetch_more_tokens() 2023-07-26 12:15:22.225784245 File "/usr/local/lib/python3.9/dist-packages/yaml/scanner.py", line 223, in fetch_more_tokens 2023-07-26 12:15:22.225785210 return self.fetch_value() 2023-07-26 12:15:22.225786212 File "/usr/local/lib/python3.9/dist-packages/yaml/scanner.py", line 577, in fetch_value 2023-07-26 12:15:22.225787070 raise ScannerError(None, None, 2023-07-26 12:15:22.225788017 yaml.scanner.ScannerError: mapping values are not allowed here 2023-07-26 12:15:22.225788879 in "<unicode string>", line 180, column 15: 2023-07-26 12:15:22.225789793 input_args: preset-rtsp-restream-low-latency 2023-07-26 12:15:22.225790624 ^ 2023-07-26 12:15:22.225791363 2023-07-26 12:15:22.225792311 ************************************************************* 2023-07-26 12:15:22.225807173 *** End Config Validation Errors *** 2023-07-26 12:15:22.225808212 ************************************************************* 2023-07-26 12:15:22.233284544 2023/07/26 12:15:22 [error] 134#134: *1 connect() failed (111: Connection refused) while connecting to upstream, client: 10.19.64.2, server: , request: "GET /api/stats HTTP/1.1", upstream: "http://127.0.0.1:5001/stats", host: "10.19.64.2:5000" 2023-07-26 12:15:22.233334360 10.19.64.2 - - [26/Jul/2023:12:15:22 +0100] "GET /api/stats HTTP/1.1" 502 157 "-" "HomeAssistant/2023.7.3 aiohttp/3.8.5 Python/3.11" "-" 2023-07-26 12:15:22.527162273 [INFO] Preparing go2rtc config... 2023-07-26 12:15:22.727177322 [INFO] The go2rtc service exited with code 1 (by signal 0) 2023-07-26 12:15:23.348278290 [INFO] Service Frigate exited with code 1 (by signal 0) 2023-07-26 12:15:23.360410949 [INFO] The go2rtc-healthcheck service exited with code 256 (by signal 15) 2023-07-26 12:15:23.395323650 [INFO] Service NGINX exited with code 0 (by signal 0) config.yml Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.