Same issue here.
I fixed it with:
pip3 install --force-reinstall aiohttp
I submitted an issue in Github
3 Likes
This is easy, use two browsers(chrome and firefox for example) , one for each instance
oh my, whats happening here:
can no longer read the template editor…
clear theme:
dark blue (ouch):
dark red:
seems the code words are always in the same color, and not set in the theme at all?
@Frenk
great update!
Is there any progress - a growing number of us are still struggling with a ongoing issue with Zigbee devices using DECONZ embedded in HA are going unavailable early morning for a few hours and sporadically becoming available till around 5am EST and it appears to be due to HA processing occurring in the morning that renders the system effecting Zigbee. However Z-wave is unaffected.
Use !secret in the configurations:
password: '!secret sambaPass'
I had that issue and solved it by entering 3000 on the host entry under the configuration of the Zwave JS Supervisor Add-on. Then I added the integration selecting the URL as ws://localhost:3000. It worked.
What a mess: Not sure if this is supervisor update or core but when I upgrade an add-on it now creates a backup of that add-on which is rather lethal in combination with the wonderful Google backup add-on - at least if you are using my settings: I keep one version on Google drive and 7 local.
So what happens when I update an add-on is the following - HA decides it needs to take a backup, Google drive add-on thinks that is the latest backup I have (which it is at that point) decides, in line with what I have asked, to delete previous cloud backup and upload the newest one.
Because of the build.yaml local add-on problem I had issues with add-ons and experimented a bit too much. Long story short, mounting the drive on another linux gives access to the partition where the local backups are but I’ve had stressful moments.
I managed to miss this breaking change again - must become better at reading breaking changes.
Just click on the switch to turn off the backup
1 Like
You just need to force close the app and open it again and it will be back to normal.
1 Like
When I want to upgrade I get the following message.
This was after I started the upgrade.
I had already created a manual snapshot.
Hmm, The interface just refreshed and it looks it just upgraded.
Screenshot?
Why?
That is probably the worst idea ever.
The other way around would make much more sense
3 Likes
Is there a way to assign an area to a bunch of entities at once? If I have to do them one at a time I’ll probably never bother!
Bumped a local add-on to a new version so it would trigger an update and then I saw it. I did see it on the core update but overlooked it on addons.
Add to to-do list: after reading breaking changes - click buttons slower and read what it says.
Thx!
True - but I have less space restrictions locally than in the cloud. I must confess I repurposed the pi that is a syncthing group with my backup folder was repurposed as an Octoprint server. So my local is taken of the server (but not when I needed it).
Looks like the latest supervisor version has some translation issues
Shortly after reboot I get these errors
Thanks! force reinstall of aiohttp worked for me too. 
As someone replied already, if you block access to version.home-assistant.io on your network then HA will never know there is an update so it can’t auto-update.
That being said, just be aware, that does not mean nothing will break. To ensure nothing will ever break you would need to essentially cut off internet access entirely for HA, any software HA depends on and any devices HA depends on. If you can’t do that then its likely some of those pieces are relying on external services and their APIs or reaching out to external services for updates. Which means at any point other developers could push an update to their service or your device which is not backwards compatible and breaks HA’s ability to communicate with it. And then you would need to update HA in order to regain that capability.
You also would not want HA externally accessible if running in this “zombie mode” as you would be ignoring potential security updates, like with the one a few weeks ago. If you are not going to update your system it really should not be exposed. Otherwise you risk exposing access to a system with known vulnerabilities since a security update was released and you didn’t know about it. Hence why I suggest cutting off all internet access to HA as well.
If you can cut off all internet access from those things are able to run in complete isolation then props. Cut off access to the HA version site and zombie it up so to speak.
2 Likes
They’re that person’s passwords, though. They have chosen them, they live with the choice.
That’s kind of what HA is all about.
There should be an option to turn this off.
1 Like
Ive seen at least two ways to do this already.
I don’t think it should be easy to turn off, but something that requires a bit of thought