can someone help me I had a problem w/ my doubletake config mqtt still reconnecting and frigate is error, and I recheck I fill in the right ip address user and password
Did you just update the frigate addon to v0.11.0? In the latest version of the add-on, the http API is disabled by default. In your Home Assistant instance, go to Settings > Add-ons > Frigate NVR > Configuration and under Network toggle on “show disabled ports”. Key in “5000” in “Web interface”. Save and restart the add-on.
2 Likes
Can someone figure out my mqtt still reconnecting which i check i have fill the right ip address user and password correct.Did someone had this problem before?
Can you verify that your mqtt broker is running? Try using mqtt explorer to see if the frigate topics are being published
yes,its running my frigate work perfectly but my doubletake still reconnecting all the time! and i dont know why
image|690x339
I’ve been following this project with great interest and wanted to say thank you for all the hard work on this. I am just starting to get my feet wet with this and had a question. For some reason Double Take doesn’t see my Frigate. MQTT is working fine but DT cannot connect to Frigate.
I get the following in the log. Frigate is working fine. I am running HASSOS on a Pi4 at the moment.
I am sure it’s something stupid but I am stumped.
error: Error: connect ECONNREFUSED 172.30.32.1:5000
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1187:16)
Answered my own question after a day of being stumped. In case someone else comes across the issue.
If you used the add-on version of frigate in HASSOS your Double Take config should look like this:
frigate:
url: http://ccab4aaf-frigate:5000
2 Likes
After some testing with double-take on home assistant I decided to offload it to a different machine with more beef. Unfortunately trying to run this as a stand-alone container has proven to be somewhat difficult. I can get it up and running and with the logs set to debug there are no errors to speak of but when I try to connect to the machine on port 3000 I just get a open and shut session.
Anyone have any thoughts as to why?
13:50:33.212903 IP 172.16.1.221.30339 > mordor.3000: Flags [S], seq 954524804, win 64240, options [mss 1460,nop,wscale 8,nop,nop,sackOK], length 0
13:50:33.213302 IP mordor.3000 > 172.16.1.221.30339: Flags [R.], seq 0, ack 1, win 0, length 0
13:50:33.518393 IP 172.16.1.221.30340 > mordor.3000: Flags [S], seq 1429972125, win 64240, options [mss 1460,nop,wscale 8,nop,nop,sackOK], length 0
13:50:33.518717 IP mordor.3000 > 172.16.1.221.30340: Flags [R.], seq 0, ack 1, win 0, length 0
13:50:33.738284 IP 172.16.1.221.30339 > mordor.3000: Flags [S], seq 954524804, win 64240, options [mss 1460,nop,wscale 8,nop,nop,sackOK], length 0
13:50:33.738603 IP mordor.3000 > 172.16.1.221.30339: Flags [R.], seq 0, ack 1, win 0, length 0
13:50:34.068117 IP 172.16.1.221.30340 > mordor.3000: Flags [S], seq 1429972125, win 64240, options [mss 1460,nop,wscale 8,nop,nop,sackOK], length 0
13:50:34.068439 IP mordor.3000 > 172.16.1.221.30340: Flags [R.], seq 0, ack 1, win 0, length 0
The traffic make it in, but I get nothing back from Double-Take.
martoq@mordor:~/docker$ telnet localhost 3000
Trying ::1...
Connected to localhost.
Escape character is '^]'.
Connection closed by foreign host.
martoq@mordor:~/docker$
And heres my docker info
martoq@mordor:~/docker$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
b089f339d7ec jakowenko/double-take "/bin/bash ./entrypo…" 2 hours ago Up 7 minutes 0.0.0.0:3000->3000/tcp, :::3000->3000/tcp double-take
And here is my config. I didn’t post my docker file because thats pretty much default,
# Double Take
# Learn more at https://github.com/jakowenko/double-take/#configuration
mqtt:
host: homeassistant.local
username: <redacted>
password: <redacted>
frigate:
url: http://homeassistant.local:5000
detectors:
deepstack:
url: http://localhost:5001
key: <redacted>
# number of seconds before the request times out and is aborted
timeout: 20
# require opencv to find a face before processing with detector
opencv_face_required: false
Regarding your previous post and the address of frigate…were you running the addon and now moved off that machine to a different host running in docker?
Yes, totally different machine and setup. My HASSOS is on Pi4 and I moved Double-Take and Deepstack to an old laptop running lubuntu 22.04.
Okay…then I think you’d need to supply the actual ip address of frigate.
Hi,
I also had trouble running double take addon. This is how I got things running.
I added repo
Dobule take
I installed double take and deep stack ( cpu ) addon from the repo.
In deep stack addon I configured port 5001
In double take addon config I use:
mqtt:
host: homeassistant:1883
username: <>
password: <>
frigate:
url: http://ccab4aaf-frigate:5000
detectors:
deepstack:
url: http://homeassistant:5001
timeout: 30
and it works basically out of the box. In homeassistant I configured my host name as homeassistant.
eats too much ram so the system hass to constantly reboot
Hello,
I am new to Home Assistant. I already have a Coral TPU. Not installed yet.
Before I spend a lot of time in research and the implementation… is it possible the following this:
I walk into the camera area in front of my garage. Double Take recognizes my face. If I do the thumbs up at the same time, an action is triggered (the garage door opens).
The main question is: is it also possible to detect a gesture (thumbs up)?
Thanks a lot
Double take doesn’t have gesture recognition.
1 Like
Does anyone have a notification automation that is triggered using the camera sensor as opposed to the user sensors? There are some benefits like
- Not needing to modify the automation to add additional users when new users are trained
- If multiple people are detected, can report all matches and non-matches in one notification instead of sending multiple… example “Billy, Alice and 1 other person were detected at the front door”
Hello everyone,
I have Nvidia Jetson Nano 4GB.
What is the best software currently supported on this platform to be used together with HA (which will run on this box) for GPU-accelerated object recognition and face detection?
Hi,
I setup double take and I’m using deepstack detector. After a few weeks I’m not that impressed with face recognition. It’s seems like it doesn’t learn well or learning method is not at it’s bets. I trained it with nearly 50 images for one person, but it doesn’t detect person correctly.
What is best practice to make detector detect person correctly?