Correct. I’ve never had to do that before didn’t know if you had a good resource off the top of your head since you just did it. If not I’ll explore.
Thanks!
It’s really easy. Go to your supervisor then snapshots. Select a snapshot from previous 2021.6.x version and click on restore
This is because there is also difference in knowledge between devs. If a higher voted FR is not developed yet it could as well mean that there are fragments not developed because the developer with that skill has some more pressing manor to attend to. It could be number of things.
Thanks for the help. Rolled it back a version but it appears the error still persists. All of my automation appear to still be working and I can create new ones… So not sure what exactly is broken or causing the error here.
Anybody have success with installing the Ambee integration. I have an APi key, but at a loss to know what I must fill in for a Name. Just keep getting invalid api key.
He doesn’t but I do like when I get a meaningful reply.
Which is not to imply I don’t get one often, I do, I just think it is nice to acknowledge it even if I don’t like the actual answer 
1 Like
I worked around this issue by pulling zwavejs2mqtt “master” instead of “latest” in my docker-compose file. Once there is a new formal release of zwavejs2mqtt, then the reco will be to switch back to “latest” to get a stable version as master could break anytime.
Thank you, this worked for me
SOLVED. For raspberry user ( core mode)
Disinstall FFMPEG
sudo apt remove ffmpeg
Compile FFMPEG from source
wget http://www.ffmpeg.org/releases/ffmpeg-4.4.tar.xz
tar xf ffmpeg-4.4.tar.xz
cd ffm*
./configure --enable-shared --enable-pic --extra-ldflags="-latomic"
make -j4
sudo make install
In venv install av
pip3 install av
modify configuration.yaml
ffmpeg:
ffmpeg_bin: /usr/local/bin/ffmpeg
bye
This update broke my duckdns - NGNIX https access. Had to revert. Anyone else? Local works.
There is a breaking change to do with trusted proxies when using a reverse proxy. This example worked for me.
Did you enter that code via the Nginx Proxy Manager gui, or in a yaml file. Thanks.
It’s in the HTTP section of the configuration.yaml file.
(I use Caddy not Nginx for reverse proxy)
JP
In the http section of your config
Seems the update to python broke tensorflow, tensorflow only supports python 3.6-3.8
2021-07-10 14:48:27 ERROR (SyncWorker_1) [homeassistant.util.package] Unable to install package tensorflow==2.3.0: ERROR: Could not find a version that satisfies the requirement tensorflow==2.3.0 (from versions: none)
Modbus Users beware with updating to 2021.7.1 if you poll multiple modbus hosts. There is an issue thats been found and currently being fixed if sensor addresses being polled are the same across multiple modbus hosts.
2 Likes
The updated MusicCast integration quietly changed the entity IDs. My device got renamed from media_player.dining_room_main to media_player.dining_room, which broke my automations and dashboard. Not a big deal but it would be good to mention this in the release notes.
Now the device has a unique ID so the entity ID can be overridden.
I have it working on my test install, I just added it in the UI and it just worked. The name for me was home and the latitude and longitude from my Install ( it just auto filled in the UI, I assume from the HA config )
Same here, and I have use_x_forwarded_for enabled and trusted_proxies configured in the http section. What am I missing?
I want to keep using http on my local network. Is that possible?
You can if you use a reverse proxy