2021.3: My Oh My

[fletch101e]

Upgrade fails with this error message:

502: Bad Gateway

Any fixes for this? Thanks

[gohm (Goh Morihara)]

I love the new UI letting me choose items, but there’s a big problem in Automations’ editor.

The new UI breaks existing Automations which uses templates to set numerical values when read. It should fall back to YAML mode, but instead, it tacitly deletes it showing in the UI mode.
It’s hard to notice the Automation you are modifying has been changed in a different part you haven’t touched.

[Shaad (David)]

So nobody has issue with recorder using MariaDB ?

[DavidFW1960 (David)]

no… but the url did change (specifically the encoding string at the end of the url changed) so that might be your issue? Can you post your URL or check the mariadb docs in the addon?

[jhein (Julian)]

Or maybe pull everything together into one main HA instance: https://github.com/custom-components/remote_homeassistant

[amelchio]

@amelchio

The numeric_state will no longer trigger when going from unavailable to within range, it must have a valid reading outside the range before it will trigger.

This change is being reverted for 2021.3.1 and unavailable numeric states will then be treated as a value outside the trigger range. That should stop the warnings.

[Faramir]

Well I had to revert back to the previous version because this update broke my Sonoff iFan03 integration…

More details:

2021.03 and tasmotized Sonoff IFan03 - lots of errors in log and fan unavailable

Hi Guys I have updated to the latest Home-Assistant 2021.03 today. I am running HA in a docker container. With this latest update my tasmotized iFan03 is now showing as unavailable. Relevant log shows: 2021-03-05 05:47:47 ERROR (MainThread) [homeassistant.util.logging] Exception in availability_message_received when handling msg on 'tele/fan1/LWT': 'Online' Traceback (most recent call last): File "/usr/src/homeassistant/homeassistant/components/mqtt/debug_info.py", line 32, in wrapper …

[Faramir]

Also… with the docker image tags… could we have docker tags for each major release that would get updated to point to the latest point release with fixes… something like the below:

Docker Tag          points to version
latest              2021.3.0
2021.3              2021.3.0
2021.2              2021.2.3

A bit like what was mentioned in:

Why isn't there a minor version docker tag e.g. 0.114.x?

Right now all home assistant releases on docker are tagged with the specific version, at the time of writing the last version would be 0.114.3. Upgrading between major releases such as 0.113.x to 0.114.x is usually associated with some non backwards compatible changes, and therefore I would say that it’s not recommended to be done automatically - when I had this set up in the past I’d often come home from work and find something isn’t working, and requires manual fixing. But upgrading between …

Having such docker tags would allow me to pin my docker image to a particular major release while still receiving point fix releases.

[djelo (Patrick Kishino)]

That would be very unlikely considering the other 20ish devices defined like this (broadlink switches) are all fine and no complaints when used as heaters.

[lolouk44]

anybody else having issues with a Xiaomi Vacuum?
I deleted the old config, rebooted, then tried to add a new integration.
After I enter the Vacuum IP and token it fails to connect
(The token was working until now, not sure why it’s failing now)
It’s a First Gen (SDJQR01RR)

[stamandr]

Smartthings integration fails after update to core-2021.3.0.

Error

Config flow could not be loaded

Logger: homeassistant.setup
Source: helpers/network.py:124
First occurred: 4:31:33 PM (1 occurrences)
Last logged: 4:31:33 PM

Anybody else having this problem ?

[Daniel_Cook (Dan Cook)]

This AM I upgraded from core-2021.2.X to core-2021.3.0, running on Hassio (host = Raspberry Pi).

Everything is behaving normally except: common add-ons like SSH & Web Terminal, HACS, and especially Node-Red seem to lose connection to the server and need to reload their page (almost) every time I come back to the relevant tab on the browser.

In other words if I have multiple Chrome tabs open, and one of those tabs is loaded up with an HA add-on, and I go to some other Chrome tab and then revisit the add-on after some time, the page does a hands-free reload with the message “Connection lost. Reconnecting…” in the bottom left. And it seems that the longer I am away from the relevant tab, the more likely the need to reconnect.

I’ve restarted the supervisor, rebooted the host RPi device, etc., none of these has stopped the effect. Node-Red is the real bummer, since trying to catch messages in the debug window is hard when it reloads every couple of minutes. I updated Chrome to the latest (89.0.4389.72) and this stopped the effect for a while, but then it started happening again. Not sure if that’s causation or just correlation. I will keep trying to pin down the root cause.

On the bright side, a Google Nest that refused for a few weeks to list to catt cast_site... calls magically started working again upon upgrade. Go figure.

[lmamakos (Louis Mamakos)]

Developers please add one simple function, let’s name it GOOD BYE or MAKE ME ZOMBIE or what ever

You might want to just run the Home Assistant container and update it manually. Sure, you lose out on all the pre-built extensions, but most can fairly easily be installed as containers manually. Learn a bit about docker-compose and you have a repeatable way to launch and manage your container stack. And manage it as deliberate actions on your part to introduce change on your schedule.

This is what I do, and I migrate to new releases usually at the .1 or .2 version, and when my wife isn’t at home (or will be asleep for a few hours…) Gotta maintain that WAF metric!

[admun (Edmond Hui)]

Interesting, I have this setting off in 2021.2,and it’s working.

Enabled it then upgrade, now works.

thanks @pepe59

[wfenwick (Will F)]

I’m getting that same entry in the logs, along with several of these when syncing entities:

2021-03-04 21:46:51 ERROR (MainThread) [homeassistant.components.google_assistant.smart_home] Error serializing light.overhead
Traceback (most recent call last):
  File "/usr/src/homeassistant/homeassistant/components/google_assistant/helpers.py", line 493, in sync_serialize
    "baseUrl": get_url(self.hass, prefer_external=True),
  File "/usr/src/homeassistant/homeassistant/helpers/network.py", line 84, in get_url
    request_host = _get_request_host()
  File "/usr/src/homeassistant/homeassistant/helpers/network.py", line 124, in _get_request_host
    raise NoURLAvailableError
homeassistant.helpers.network.NoURLAvailableError

[Daniel_Cook (Dan Cook)]

NB: updating to 2021.3.1 didn’t fix the behavior.

[tom_l]

@Mariusthvdb

the system is very much less responsive

I wouldn’t say “very much” but the frontend it is noticeably slower to load and drops the connection much more readily (firefox 86.0 and iOS app). No noticeable change in resource use or issues in the browser inspector.

[reotto (Rick Otto)]

pfSense has this figured out. You update the core, and it goes to a system will be available in 90 seconds message and counts down. If it isn’t back at 90, it adds another 20 until it is available again. This can be figured out.

[solidssss]

Device_tracker is not wokring with either core-2021.3 or core-2021.3.1. Just says “unknown” on all devices.

[Shaad (David)]

This is the URL I use in configuration.yaml in the recorder section :

  db_url: mysql://hass:xxxxxxxxxxx@core-mariadb/homeassistant?charset=utf8