[SUPPORT] blakeblackshear - Frigate


Recommended Posts

38 minutes ago, DanielWWS said:

using unraid and seem to be getting a constant error popping up at every detection: 

 

 

 

##  the start up process looks okay to me but im a novice.  below log from start up ## 

 

 

this is my config.yml file:

 

 

Ive installed the coral drivers but the TPU's are USB's and ive tried just one or the other and we get the same results.  Its when a detection happens this starts.  The usb's are both in USB3 ports and its got 32gb ddr3 memory // I7-4960x@3.60GHz processor and ive built this machine pretty much just to test and run this (from old parts but its all good.) 

 

Version: DEBUG 0.9.1-800F33E

 

coral1 = inference speed - 10

coral2 = inference speed - 29.9

 

Does anyone have any ideas how to stop this reoccuring error?  can it be added for someone to look into dealing with it if its a bug in unraid? 

 

 

I'm out for Holidays till tuesday so I can't help you but just for helping a little, the usb coral does not need the drivers so you can uninstall them. Make sure that you are binding the correct paths to the folders and that the container is on privileged mode.

  • Like 1
Link to comment
I'm out for Holidays till tuesday so I can't help you but just for helping a little, the usb coral does not need the drivers so you can uninstall them. Make sure that you are binding the correct paths to the folders and that the container is on privileged mode.

Hi, thank you and I hope you enjoy your holidays. No mad rush! We’ve all got to have some down time.

I installed the coral drivers after the problem started. Pushed a restart and hoped it would deal with the problem but it didn’t. Just deleted them and will test and report back before Tuesday.

All other settings not changed. 7000f1eef9a8ab659415dbf4069babc9.jpg


Sent from my iPhone using Tapatalk
Link to comment

Hi,

 

having the following error since I updated to 0.9

Fatal Python error: Fatal Python error: Bus error

 

When I looked for some solution, I figured out it's probably related to the shm-size like explained here

 

So I added the following variable into my docker template but it didn't solve the issue. 

After 10 to 60min, the error pops up in the log and my camera turn white.

 

I am pretty sure this shm-size can't be set for each docker container and should be setup for Docker itself but I have no idea how to do that.

 

Any idea?

Thanks

 

image.png

Edited by mathgoy
Link to comment

So this may not be the correct place for this, but since this is the thread that for the coral drivers, I'd thought I would give this a shot...

 

Anybody have an idea on what I need to do to get the dual edge TPU to even show up in the OS? I would expect to see one device listed in lspci, but I see nothing unfortunately. I have verified that the wifi e-key slot is enabled in the BIOS. Any ideas on where to go from here would be helpful.

 

Motherboard is an AsRock H570 Steel Legend with an Intel 10700K CPU (don't mind the K, it was part of a bundle and I had it lying around).

 

Thanks

Link to comment
15 hours ago, mathgoy said:

Hi,

 

having the following error since I updated to 0.9

Fatal Python error: Fatal Python error: Bus error

 

When I looked for some solution, I figured out it's probably related to the shm-size like explained here

 

So I added the following variable into my docker template but it didn't solve the issue. 

After 10 to 60min, the error pops up in the log and my camera turn white.

 

I am pretty sure this shm-size can't be set for each docker container and should be setup for Docker itself but I have no idea how to do that.

 

Any idea?

Thanks

 

image.png

 

 

So it looks like I found the solution.

 

In the Extra Parameters of the Docker template, I added --shm-size=1G

 

So the overall Extra Parameters field looks like:

--runtime=nvidia --shm-size=1G

 

Link to comment
Just now, mathgoy said:

 

 

So it looks like I found the solution.

 

In the Extra Parameters of the Docker template, I added --shm-size=1G

 

So the overall Extra Parameters field looks like:

--runtime=nvidia --shm-size=1G

 

My bad, I deleted that but as I see on the changelog is still needed. I will add it again tomorrow.

Link to comment
On 10/10/2021 at 9:17 PM, SMLLR said:

So this may not be the correct place for this, but since this is the thread that for the coral drivers, I'd thought I would give this a shot...

 

Anybody have an idea on what I need to do to get the dual edge TPU to even show up in the OS? I would expect to see one device listed in lspci, but I see nothing unfortunately. I have verified that the wifi e-key slot is enabled in the BIOS. Any ideas on where to go from here would be helpful.

 

Motherboard is an AsRock H570 Steel Legend with an Intel 10700K CPU (don't mind the K, it was part of a bundle and I had it lying around).

 

Thanks

@ich777 any idea?

Link to comment
On 10/8/2021 at 7:00 PM, DanielWWS said:

using unraid and seem to be getting a constant error popping up at every detection: 

 

 

 

##  the start up process looks okay to me but im a novice.  below log from start up ## 

 

 

this is my config.yml file:

 

 

Ive installed the coral drivers but the TPU's are USB's and ive tried just one or the other and we get the same results.  Its when a detection happens this starts.  The usb's are both in USB3 ports and its got 32gb ddr3 memory // I7-4960x@3.60GHz processor and ive built this machine pretty much just to test and run this (from old parts but its all good.) 

 

Version: DEBUG 0.9.1-800F33E

 

coral1 = inference speed - 10

coral2 = inference speed - 29.9

 

Does anyone have any ideas how to stop this reoccuring error?  can it be added for someone to look into dealing with it if its a bug in unraid? 

 

 

Can you test your Coral doing this on another machine like a raspberry, a Windows or a VM?:

 

https://coral.ai/docs/accelerator/get-started/

 

Link to comment
On 10/10/2021 at 9:17 PM, SMLLR said:

Anybody have an idea on what I need to do to get the dual edge TPU to even show up in the OS? I would expect to see one device listed in lspci, but I see nothing unfortunately. I have verified that the wifi e-key slot is enabled in the BIOS. Any ideas on where to go from here would be helpful.

@yayitazale maybe... :D

Maybe the slot you are putting it in is not fully compatible with the PCIe specification since the Dual Edge TPU makes use of two PCIe 2.0 x1 lanes and most M.2 E-Keyed slots are not "really" compatible with the PCIe specifications.

 

I also learned that the hard way because I also needed an adapter for a PCIe slot that are really hard to find and now only one TPU is working because I only got a single PCIe x1 Slot free, but better than nothing...

Here is an entire issue on Github about what is compatible: Click

 

My recommendation would be to stick with a A+E-Key Coral TPU or with a B+M-Key Coral TPU, these are some weird devices or at least the compatibility is a littel weird... :D

 

An extra question from my side: may I ask how do you cool the Dual Edge TPU, since this needs extra cooling.

  • Like 1
Link to comment
5 hours ago, takkkkkkk said:

are there anyways to know if my GPU acceleration is working? camera feeds are working, but when I don't see any processes when I run "nvidia-smi" in the terminal (I do see drivers wattage info, etc). anythoughts on what I can do to troubleshoot?

Run nvidia-smi on the Unraid console not the containers one.

  • Like 1
Link to comment

For those who updated to V0.9.2 from 0.9.1 and found their config.yml file was no longer valid....

 

In V0.9.1, I didn't have RTMP included in my config.yml and I would get an error "Camera (camera name) has rtmp enabled, but rtmp is not assigned to an input."  but it would still work fine.

 

With the change to V0.9.2, it no longer works - lists it as a configuration error and shuts down the docker.  

 

Under each camera section, I added the lines below that seemed to fix it.  I no longer get the error in my logs and frigate is working fine again.  "rtmp:" should line up with your "ffmpeg", "detect", "record" etc.  You'll need the below lines for every camera.

    rtmp:
      enabled: false 

 

Here is the error message I was getting in my logs - again - solved with the above yml code.

 

Camera deck has rtmp enabled, but rtmp is not assigned to an input. (type=value_error)
*************************************************************
*** 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.

  • Like 1
Link to comment

I've also noticed that the default recording format has changed so if you have explicit output_args defined you may want to review them (changed from mp4 to ts)

 

Audio is now recording by default - the entire reason I had a output_args ffmpeg override anyway.


RRB

  • Like 1
Link to comment

I have updated to the latest version, made sure my config is correct (at least, I can't find any faults in it). But when I start the container it stops again within a minute or so. The logs give me this:

[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.
[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.
[2021-10-22 15:07:19] frigate.app INFO : Starting Frigate (0.9.2-25bb515)
Starting migrations
[2021-10-22 15:07:19] peewee_migrate INFO : Starting migrations
There is nothing to migrate
[2021-10-22 15:07:19] peewee_migrate INFO : There is nothing to migrate
[Errno -2] Name or service not known
[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.

 

My Config is as follows: (removed credentials and stuff offcourse):

mqtt:
  host: homeassistant.fqdn.local
  port: 1883
  topic_prefix: blaat
  client_id: blaat
  user: blaat
  password: blaat
  stats_interval: 60

detectors:
  coral:
    type: edgetpu
    device: usb
    num_threads: 3

database:
  path: /media/frigate/frigate.db

birdseye:
  enabled: True
  width: 640
  height: 480
  quality: 8
  mode: motion

ffmpeg:
  input_args:
    - -avoid_negative_ts
    - make_zero
    - -fflags
    - nobuffer
    - -flags
    - low_delay
    - -strict
    - experimental
    - -fflags
    - +genpts+discardcorrupt
#    - -rw_timeout
#    - "5000000"
    - -use_wallclock_as_timestamps
    - "1"

detect:
  width: 1280
  height: 720
  fps: 5
  enabled: True
  max_disappeared: 25

objects:
  track:
    - person

motion:
  threshold: 25
  contour_area: 100
  delta_alpha: 0.2
  frame_alpha: 0.2
  frame_height: 180

record:
  enabled: False
  retain_days: 0
  events:
    max_seconds: 300
    pre_capture: 5
    post_capture: 5
    objects:
      - person
    required_zones: []
    retain:
      default: 10
      objects:
        person: 15

snapshots:
  enabled: true
  timestamp: false
  bounding_box: False
  crop: False
  height: 175
  retain:
    default: 10
    objects:
      person: 15

rtmp:
  enabled: False

cameras:
  reolink-test:
    ffmpeg:
      inputs:
#        - path: rtsp://user:pass@192.168.103.141:554/h265Preview_01_main
#          roles:
#            - record
#            - clips
#            - detect
        - path: rtsp://user:pass@192.168.103.141:554//h264Preview_01_sub
          roles:
            - record
            - detect
#            - clips
    best_image_timeout: 60

Does anyone have an idea what's wrong?

Link to comment
2 minutes ago, Guido said:

I have updated to the latest version, made sure my config is correct (at least, I can't find any faults in it). But when I start the container it stops again within a minute or so. The logs give me this:

[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.
[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.
[2021-10-22 15:07:19] frigate.app INFO : Starting Frigate (0.9.2-25bb515)
Starting migrations
[2021-10-22 15:07:19] peewee_migrate INFO : Starting migrations
There is nothing to migrate
[2021-10-22 15:07:19] peewee_migrate INFO : There is nothing to migrate
[Errno -2] Name or service not known
[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.

 

My Config is as follows: (removed credentials and stuff offcourse):

mqtt:
  host: homeassistant.fqdn.local
  port: 1883
  topic_prefix: blaat
  client_id: blaat
  user: blaat
  password: blaat
  stats_interval: 60

detectors:
  coral:
    type: edgetpu
    device: usb
    num_threads: 3

database:
  path: /media/frigate/frigate.db

birdseye:
  enabled: True
  width: 640
  height: 480
  quality: 8
  mode: motion

ffmpeg:
  input_args:
    - -avoid_negative_ts
    - make_zero
    - -fflags
    - nobuffer
    - -flags
    - low_delay
    - -strict
    - experimental
    - -fflags
    - +genpts+discardcorrupt
#    - -rw_timeout
#    - "5000000"
    - -use_wallclock_as_timestamps
    - "1"

detect:
  width: 1280
  height: 720
  fps: 5
  enabled: True
  max_disappeared: 25

objects:
  track:
    - person

motion:
  threshold: 25
  contour_area: 100
  delta_alpha: 0.2
  frame_alpha: 0.2
  frame_height: 180

record:
  enabled: False
  retain_days: 0
  events:
    max_seconds: 300
    pre_capture: 5
    post_capture: 5
    objects:
      - person
    required_zones: []
    retain:
      default: 10
      objects:
        person: 15

snapshots:
  enabled: true
  timestamp: false
  bounding_box: False
  crop: False
  height: 175
  retain:
    default: 10
    objects:
      person: 15

rtmp:
  enabled: False

cameras:
  reolink-test:
    ffmpeg:
      inputs:
#        - path: rtsp://user:pass@192.168.103.141:554/h265Preview_01_main
#          roles:
#            - record
#            - clips
#            - detect
        - path: rtsp://user:pass@192.168.103.141:554//h264Preview_01_sub
          roles:
            - record
            - detect
#            - clips
    best_image_timeout: 60

Does anyone have an idea what's wrong?

Can you attach a screenshot of your container deploy?

Link to comment
2 hours ago, yayitazale said:

Just delete the container and the template image and launch it again with the new template at it will work.

Did that just now, but it didn't help. Same message.

 

And I did see it pulling the entire image again, so that's not the issue.

Edited by Guido
Link to comment

I had it running once... but after a reboot it didn't start again and I have the same error message in the log.

 

I hope anyone has another idea, cause I would really like to get Frigate running again.

Edited by Guido
Link to comment
1 hour ago, B0rax said:

I have the same issue with a complete new install. The docker just terminates with the same error as yours "Name or service not known"

 

12 hours ago, Guido said:

I had it running once... but after a reboot it didn't start again and I have the same error message in the log.

 

I hope anyone has another idea, cause I would really like to get Frigate running again.

 

As far as I read on these closed issues on github, it looks like a MQTT connection problem. Please use the IP instead of the FQDN.

 

https://github.com/blakeblackshear/frigate/issues/2066

https://github.com/blakeblackshear/frigate/issues/2094

 

If that doesn't works please open an issue on github, this doesn't look like a an error related with Unraid template.

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.