[SUPPORT] blakeblackshear - Frigate


Recommended Posts

1 hour ago, CryPt00n said:

Logs are attached

First of all it is still bound to VFIO as you can see from the Kernel driver in use:

01:00.0 System peripheral [0880]: Global Unichip Corp. Coral Edge TPU [1ac1:089a]
    Subsystem: Global Unichip Corp. Coral Edge TPU [1ac1:089a]
    Kernel driver in use: vfio-pci
    Kernel modules: apex

 

That's mainly caused because you have a file named "vfio-pci - Kopie.cfg" in your /boot/config directory with this content:

BIND=0000:01:00.0|10de:11c0 0000:01:00.1|10de:0e0b

I would strongly recommend to remove both vifo files from the /boot/config directory and reboot.

 

After that it should work just fine, if you changed the mapping in the template already like:

grafik.png.3fb01e2a56a0d8e6ab6c32c8bc8e83a3.png

Link to comment

 

30 minutes ago, ich777 said:

First of all it is still bound to VFIO as you can see from the Kernel driver in use:

01:00.0 System peripheral [0880]: Global Unichip Corp. Coral Edge TPU [1ac1:089a]
    Subsystem: Global Unichip Corp. Coral Edge TPU [1ac1:089a]
    Kernel driver in use: vfio-pci
    Kernel modules: apex

 

That's mainly caused because you have a file named "vfio-pci - Kopie.cfg" in your /boot/config directory with this content:

BIND=0000:01:00.0|10de:11c0 0000:01:00.1|10de:0e0b

I would strongly recommend to remove both vifo files from the /boot/config directory and reboot.

 

After that it should work just fine, if you changed the mapping in the template already like:

grafik.png.3fb01e2a56a0d8e6ab6c32c8bc8e83a3.png

 

 

Have deleted both vfio and the vfio bak file now, tpu is still not showing up in the plugin.

 

I have attached new logs.

server-diagnostics-20211126-1944.zip

Link to comment
1 hour ago, CryPt00n said:

Have deleted both vfio and the vfio bak file now, tpu is still not showing up in the plugin.

First your server enables the TPU after the plugin installation is finished:

Nov 26 19:40:25 server kernel: apex 0000:01:00.0: enabling device (0000 -> 0002)

 

A little while after that your server does this:

Nov 26 19:41:00 server kernel: vfio-pci 0000:01:00.0: vfio_ecap_init: hiding ecap 0x1e@0x110

 

It seems like you've bound it to your NicehashOS too and that's why it's bound again to VFIO... :D

 

-device vfio-pci,host=0000:01:00.0,id=hostdev0,bus=pci.4,addr=0x0 \

 

 

I think something is really miss configured that always binds the card to VFIO:

01:00.0 System peripheral [0880]: Global Unichip Corp. Coral Edge TPU [1ac1:089a]
	Subsystem: Global Unichip Corp. Coral Edge TPU [1ac1:089a]
	Kernel driver in use: vfio-pci
	Kernel modules: apex

 

You can always check which modules are loaded from a unRAID terminal with the following command:

lspci -k

 

  • Like 1
Link to comment
  • 2 weeks later...

I setup Frigate with my Reolink 511, and it worked awesome out of the gate.  But then, I felt saucy & moved to the 10-beta.  Now, I can't get the UI to pull up from the Unraid Docker.  No issue with the HA add-on...Not sure if it is my docker config or something I changed in either the BIOS (Asrock x570du - Changed default video from NVIDIA back to IPMI) or Unraid networking (toggled bridge mode off and back on). 

All other Unraid dockers have no issue, only Frigate.  Same time, my previously working HA add-on Frigate Proxy went down with this error - 2021/12/04 23:26:11 [error] 352#352: *219 upstream prematurely closed connection while reading upstream, client: 172.30.X.Y, server: ccab4aaf-frigate-proxy, request: "GET / HTTP/1.1", upstream: "http://unraid-IP:5000/", host: "172.30.A.B:5000"

I've deleted the docker folders & images, and tried multiple times again. 

 

I know it's working since it's constantly creating videos ~5MB videos. 

s6-init] making user provided files available at /var/run/s6/etc...exited 0.
[s6-init] ensuring user provided files have correct perms...exited 0.
[fix-attrs.d] applying ownership & permissions fixes...
[fix-attrs.d] done.
[cont-init.d] executing container initialization scripts...
[cont-init.d] done.
[services.d] starting services
[services.d] done.
[2021-12-04 23:14:24] frigate.app INFO : Starting Frigate (0.9.4-26ae608)
Starting migrations
[2021-12-04 23:14:24] peewee_migrate INFO : Starting migrations
There is nothing to migrate
[2021-12-04 23:14:24] peewee_migrate INFO : There is nothing to migrate
[2021-12-04 23:14:24] frigate.mqtt INFO : MQTT connected
[2021-12-04 23:14:24] detector.coral INFO : Starting detection process: 257
[2021-12-04 23:14:24] frigate.app INFO : Output process started: 259
[2021-12-04 23:14:24] ws4py INFO : Using epoll
[2021-12-04 23:14:24] frigate.edgetpu INFO : Attempting to load TPU as usb
[2021-12-04 23:14:24] frigate.app INFO : Camera processor started for driveway-right: 262
[2021-12-04 23:14:24] frigate.app INFO : Capture process started for driveway-right: 265
[2021-12-04 23:14:27] frigate.edgetpu INFO : TPU found
[2021-12-04 23:14:24] ws4py INFO : Using epoll
[2021-12-04 23:14:29] frigate.record WARNING : Discarding a corrupt recording segment: driveway-right-20211204231407.mp4
[2021-12-04 23:15:03] frigate.mqtt INFO : MQTT connected

Edited by OneMeanRabbit
Link to comment

I'm not sure if this is the correct place for this question, so let me know if there is a better place to post.

 

I've been using the Frigate docker container successfully on my Unraid server for many months with the Google Coral (USB Accelerator) and it has worked very well.

 

A day or two ago I noticed that events weren't being registered in Frigate.  I investigated, and it looks like my USB Coral is no longer setup correctly in Unraid.

 

It looks like previously my Coral was passed through under /dev/bus/usb and  /dev/dri/ (at least that is what was in the template for the container).  Now I do not see anything under /dev/dri/.  I did a little research and it looks like it might need to be under /dev/apex* but I do not have anything like that on my server.

 

/bin/ls: cannot access '/dev/apex*': No such file or directory

 

I uninstalled the plugin "Coral Accelerator Module Drivers", rebooted my server, installed the plugin again, rebooted, but I do not see any evidence that the plugin is recognizing my Coral.  When I do lsusb, I see it:

Bus 003 Device 003: ID 18d1:9302 Google Inc.

 

In the Unraid GUI, if I go to Settings->Coral Driver, it only shows me the heading "Coral TPU Driver" but other than that, the page is completely blank with no statistics about by Coral.

 

I've attached my Diagnostics file to this post.  I would appreciate another set of eyes on this because I am having a hard time figuring this one out and my family is getting tired of all of my reboots :)

 

Thanks in advance!

storage-diagnostics-20211206-1527.zip

Link to comment
2 hours ago, furiousgeorge said:

I'm not sure if this is the correct place for this question, so let me know if there is a better place to post.

 

I've been using the Frigate docker container successfully on my Unraid server for many months with the Google Coral (USB Accelerator) and it has worked very well.

 

A day or two ago I noticed that events weren't being registered in Frigate.  I investigated, and it looks like my USB Coral is no longer setup correctly in Unraid.

 

It looks like previously my Coral was passed through under /dev/bus/usb and  /dev/dri/ (at least that is what was in the template for the container).  Now I do not see anything under /dev/dri/.  I did a little research and it looks like it might need to be under /dev/apex* but I do not have anything like that on my server.

 

/bin/ls: cannot access '/dev/apex*': No such file or directory

 

I uninstalled the plugin "Coral Accelerator Module Drivers", rebooted my server, installed the plugin again, rebooted, but I do not see any evidence that the plugin is recognizing my Coral.  When I do lsusb, I see it:

Bus 003 Device 003: ID 18d1:9302 Google Inc.

 

In the Unraid GUI, if I go to Settings->Coral Driver, it only shows me the heading "Coral TPU Driver" but other than that, the page is completely blank with no statistics about by Coral.

 

I've attached my Diagnostics file to this post.  I would appreciate another set of eyes on this because I am having a hard time figuring this one out and my family is getting tired of all of my reboots :)

 

Thanks in advance!

storage-diagnostics-20211206-1527.zip 153.16 kB · 0 downloads

Same thing happened to me, I solved it by changing the docker UI from "/dev/bus/usb" to "/dev/bus/usb/003/003" in your case.  Hope this helps, mine finds it and records - just lost all UI around the same time...

  • Thanks 1
Link to comment
On 12/5/2021 at 6:28 AM, OneMeanRabbit said:

I setup Frigate with my Reolink 511, and it worked awesome out of the gate.  But then, I felt saucy & moved to the 10-beta.  Now, I can't get the UI to pull up from the Unraid Docker.  No issue with the HA add-on...Not sure if it is my docker config or something I changed in either the BIOS (Asrock x570du - Changed default video from NVIDIA back to IPMI) or Unraid networking (toggled bridge mode off and back on). 

All other Unraid dockers have no issue, only Frigate.  Same time, my previously working HA add-on Frigate Proxy went down with this error - 2021/12/04 23:26:11 [error] 352#352: *219 upstream prematurely closed connection while reading upstream, client: 172.30.X.Y, server: ccab4aaf-frigate-proxy, request: "GET / HTTP/1.1", upstream: "http://unraid-IP:5000/", host: "172.30.A.B:5000"

I've deleted the docker folders & images, and tried multiple times again. 

 

I know it's working since it's constantly creating videos ~5MB videos. 

s6-init] making user provided files available at /var/run/s6/etc...exited 0.
[s6-init] ensuring user provided files have correct perms...exited 0.
[fix-attrs.d] applying ownership & permissions fixes...
[fix-attrs.d] done.
[cont-init.d] executing container initialization scripts...
[cont-init.d] done.
[services.d] starting services
[services.d] done.
[2021-12-04 23:14:24] frigate.app INFO : Starting Frigate (0.9.4-26ae608)
Starting migrations
[2021-12-04 23:14:24] peewee_migrate INFO : Starting migrations
There is nothing to migrate
[2021-12-04 23:14:24] peewee_migrate INFO : There is nothing to migrate
[2021-12-04 23:14:24] frigate.mqtt INFO : MQTT connected
[2021-12-04 23:14:24] detector.coral INFO : Starting detection process: 257
[2021-12-04 23:14:24] frigate.app INFO : Output process started: 259
[2021-12-04 23:14:24] ws4py INFO : Using epoll
[2021-12-04 23:14:24] frigate.edgetpu INFO : Attempting to load TPU as usb
[2021-12-04 23:14:24] frigate.app INFO : Camera processor started for driveway-right: 262
[2021-12-04 23:14:24] frigate.app INFO : Capture process started for driveway-right: 265
[2021-12-04 23:14:27] frigate.edgetpu INFO : TPU found
[2021-12-04 23:14:24] ws4py INFO : Using epoll
[2021-12-04 23:14:29] frigate.record WARNING : Discarding a corrupt recording segment: driveway-right-20211204231407.mp4
[2021-12-04 23:15:03] frigate.mqtt INFO : MQTT connected

If beta is not working go back to stable. Anyway, this looks something regarding to your configuration or docker deployment. I'm running 10 beta 4 for past 2 weeks with no issues at all.

Link to comment
21 hours ago, furiousgeorge said:

I'm not sure if this is the correct place for this question, so let me know if there is a better place to post.

 

I've been using the Frigate docker container successfully on my Unraid server for many months with the Google Coral (USB Accelerator) and it has worked very well.

 

A day or two ago I noticed that events weren't being registered in Frigate.  I investigated, and it looks like my USB Coral is no longer setup correctly in Unraid.

 

It looks like previously my Coral was passed through under /dev/bus/usb and  /dev/dri/ (at least that is what was in the template for the container).  Now I do not see anything under /dev/dri/.  I did a little research and it looks like it might need to be under /dev/apex* but I do not have anything like that on my server.

 

/bin/ls: cannot access '/dev/apex*': No such file or directory

 

I uninstalled the plugin "Coral Accelerator Module Drivers", rebooted my server, installed the plugin again, rebooted, but I do not see any evidence that the plugin is recognizing my Coral.  When I do lsusb, I see it:

Bus 003 Device 003: ID 18d1:9302 Google Inc.

 

In the Unraid GUI, if I go to Settings->Coral Driver, it only shows me the heading "Coral TPU Driver" but other than that, the page is completely blank with no statistics about by Coral.

 

I've attached my Diagnostics file to this post.  I would appreciate another set of eyes on this because I am having a hard time figuring this one out and my family is getting tired of all of my reboots :)

 

Thanks in advance!

storage-diagnostics-20211206-1527.zip 153.16 kB · 0 downloads

For USB coral you don't need the Driver pluging, that only works for PCI and M.2 coral cards. It looks you have changed the config file of your frigate. Make sure you have these:

detectors:
  coral:
    type: edgetpu
    device: usb

 

And not these:

detectors:
  coral:
    type: edgetpu
    device: pci

 

The USB tpu mapping is /dev/bus/usb. /dev/dri/ is mapped for hardware acc. of intel CPUs.

Link to comment
2 hours ago, furiousgeorge said:

Wow - that fixed it!  Thank you so much for the help!

 

18 hours ago, OneMeanRabbit said:

Same thing happened to me, I solved it by changing the docker UI from "/dev/bus/usb" to "/dev/bus/usb/003/003" in your case.  Hope this helps, mine finds it and records - just lost all UI around the same time...

 

Be aware that pointing to an specific usb port will not work if you reboot your Unraid machine as port may change to another number. Thast why the mapping is done to the entire USB driver to let frigate find it whatever the port number is.

Edited by yayitazale
  • Like 1
Link to comment
18 hours ago, mikey6283 said:

Now that Frigate is running in the dockers satisfactory at the moment . How can you install the Frigate integration addon to use the  cameras motion ,detection options  to turn on switches/devices mainly on one of my cameras?  Hacs is install on VM for HA.

 Any help would be appreciated.

Google is your friend, here's a walk-through, github with docs straight from the man, & Home Assistant's community for further help with Frigate integration/Frigate Proxy/HACS. https://smarthomepursuits.com/how-to-install-frigate-in-home-assistant-for-pt1/
https://github.com/blakeblackshear/frigate-hass-addons

https://community.home-assistant.io/c/third-party/15
 

Link to comment

Brand new user - would like to try this docker out for my CCTV/NVR needs.

 

I am unable to try to see/edit/verify the /config/config.yml file as the whole docker stops and so getting to terminal won't work because the session closes.

 

How do I troubleshoot this?

 

Basic install of the Nvidia version of the docker.  Only thing I edited was changing media path to an unassigned disk out of the array, added the GPU ID for Nvidia card and password.   When I run it it auto-closes with logs showing: 

 

 

-----------

[s6-init] making user provided files available at /var/run/s6/etc...exited 0.
[s6-init] ensuring user provided files have correct perms...exited 0.
[fix-attrs.d] applying ownership & permissions fixes...
[fix-attrs.d] done.
[cont-init.d] executing container initialization scripts...
[cont-init.d] done.
[services.d] starting services
[services.d] done.
*************************************************************
*************************************************************
*** Your config file is not valid! ***
*** Please check the docs at ***
*** https://docs.frigate.video/configuration/index ***
*************************************************************
*************************************************************
*** Config Validation Errors ***
*************************************************************
[Errno 2] No such file or directory: '/config/config.yml'
*************************************************************
*** End Config Validation Errors ***
*************************************************************
[cmd] python3 exited 1
[cont-finish.d] executing container finish scripts...
[cont-finish.d] done.
[s6-finish] waiting for services.
[s6-finish] sending all processes the TERM signal.
[s6-finish] sending all processes the KILL signal and exiting.

 

 

 

Link to comment
2 minutes ago, Ystebad said:

Brand new user - would like to try this docker out for my CCTV/NVR needs.

 

I am unable to try to see/edit/verify the /config/config.yml file as the whole docker stops and so getting to terminal won't work because the session closes.

 

How do I troubleshoot this?

 

Basic install of the Nvidia version of the docker.  Only thing I edited was changing media path to an unassigned disk out of the array, added the GPU ID for Nvidia card and password.   When I run it it auto-closes with logs showing: 

 

 

-----------

[s6-init] making user provided files available at /var/run/s6/etc...exited 0.
[s6-init] ensuring user provided files have correct perms...exited 0.
[fix-attrs.d] applying ownership & permissions fixes...
[fix-attrs.d] done.
[cont-init.d] executing container initialization scripts...
[cont-init.d] done.
[services.d] starting services
[services.d] done.
*************************************************************
*************************************************************
*** Your config file is not valid! ***
*** Please check the docs at ***
*** https://docs.frigate.video/configuration/index ***
*************************************************************
*************************************************************
*** Config Validation Errors ***
*************************************************************
[Errno 2] No such file or directory: '/config/config.yml'
*************************************************************
*** End Config Validation Errors ***
*************************************************************
[cmd] python3 exited 1
[cont-finish.d] executing container finish scripts...
[cont-finish.d] done.
[s6-finish] waiting for services.
[s6-finish] sending all processes the TERM signal.
[s6-finish] sending all processes the KILL signal and exiting.

 

 

 

You have to create and place a config file, the APP does not create one. Follow the manual/docs from here: https://docs.frigate.video/

Link to comment
  • 2 weeks later...

I’ve got a couple of Reolink 410 cameras on order. Once they arrive I’ll be installing and setting up the Frigate-Nvidia docker for use with my Home Assistant vm. It’s mostly temporary since I’ve also got a Home Assistant Yellow (formerly Amber) on order and will be migrating over to it once they start shipping. Will probably also get the Coral USB dongle then.

 

I’m curious if it’s possible to migrate the Frigate configuration to another machine or are the settings different enough that it’s better to just generate a new yml file?

Edited by wgstarks
Typo
Link to comment
52 minutes ago, wgstarks said:

I’ve got a couple of Reolink 410 cameras on order. Once they arrive I’ll be installing and setting up the Frigate-Nvidia docker for use with my Home Assistant vm. It’s mostly temporary since I’ve also got a Home Assistant Yellow (formerly Amber) on order and will be migrating over to it once they start shipping. Will probably also get the Coral USB dongle then.

 

I’m curious if it’s possible to migrate the Frigate configuration to another machine or are the settings different enough that it’s better to just generate a new yml file?

You will only need to change the ffmpeg parameters according to the  documentation (and also the detectors parameters if you change the coral type) as this runs in a docker container.

 

 

Link to comment

Hi, I am working with the frigate docker container trying to use my amd apu for ffmpeg hardware acceleration. The docker container prefills the expected location of the gpu as `/dev/dri/renderD128` but `/dev/dri` does not exist. I do see a `/dev/vfio` but I am not sure if that is correct? Any help is greatly appreciated. Thanks!

Quick note: Not currently using the igpu for anything else

Link to comment
On 12/30/2021 at 8:24 PM, crzynik said:

Hi, I am working with the frigate docker container trying to use my amd apu for ffmpeg hardware acceleration. The docker container prefills the expected location of the gpu as `/dev/dri/renderD128` but `/dev/dri` does not exist. I do see a `/dev/vfio` but I am not sure if that is correct? Any help is greatly appreciated. Thanks!

Quick note: Not currently using the igpu for anything else

Is this supported?

 

https://docs.frigate.video/configuration/hardware_acceleration

  • Like 1
Link to comment

I'm having a weird issue with this docker where if the docker is started, I can't start up my Windows VM. I'm not sure what the issue is.

 

I'm using the nvidia docker passing in my Quadro and my Coral USB device. My Windows VM is using a GTX 1070 with its own PCIe USB card passed in. Home Assistant is being ran on an actual RPi 4. What could be causing this issue?

Link to comment

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.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.