I’m a fairly minimal user. I’ve got a handful of smart lights and gadgets and have set them up how I like and haven’t had to fidget or make changes in about a year. I only restart the server maybe every 3-4 months when things get really slow.
Wanted to add my lawn sprinklers back into my groups for summer and tried to change my group tab names to icons which didn’t work (all devices/browsers weren’t displaying icons). Running v0.84 I think, and there was a hass.io upgrade available in the config menu so I clicked “upgrade”. The spinning wheel sat there for 20-30min and I got bored, clicked it again a few times and it just turned green with no spinning wheel.
Tried restarting the server which worked fine, booted up again. Tried the “upgrade” button and just turned green again. A few minutes later, all my devices lost connection with the server suddenly and now I have nothing working.
Hardware is Raspberry Pi 3B+, I power cycled it and samba share comes up on my network… I can read/change config files. I can ping the IP address no problem.
There is no response from the android app. If I try DUCKDNS access I get “Corrupted Content Error” in firefox. If I try the local IP address (https://192.168.1.114:8123) all I get is “Problem Loading Page”
I read the legacy API password went away so I changed that in config files and power cycled again but still nothing…
Did I brick this thing? Is pressing the upgrade button in the UI really that risky?
Also - I am copying over all my config folder contents now to a backup location so I can inevitably spend all day setting up a new hassio install and there is 12GB worth of files…
This seem abnormal, I have all of maybe 20 devices and 30-40 automations, no special configurations or custom UIs, just a very plain install with DuckDNS and Samba as my only plugins.
My config files should be in the MB range, not 12GB…?
A. Read the log file and find out why it’s not booting. It will have your answers.
B. The file you have highlighted is the database file, as indicated by the file type. It has nothing to do with your config. It’s a running history of events and states from your HA. You can minimize it’s output by configuring your recording: component.
Thanks, I was pretty knowledgeable about this stuff a year ago but have since forgotten everything.
The home-assistant.log file?
All it contains is:
2019-07-04 13:25:10 ERROR (MainThread) [homeassistant.config] Invalid config for [homeassistant]: [switch._automatic_watering] is an invalid option for [homeassistant]. Check: homeassistant->customize->switch._automatic_watering. (See /config/configuration.yaml, line 2).
2019-07-04 13:25:10 ERROR (MainThread) [homeassistant.core] Error doing job: Task exception was never retrieved
Traceback (most recent call last):
File “/usr/src/homeassistant/homeassistant/core.py”, line 1127, in async_call
raise ServiceNotFound(domain, service) from None
homeassistant.exceptions.ServiceNotFound: Unable to find service persistent_notification/create
I haven’t actually changed anything in configuration.yaml in 7 months so I don’t know what specifically to fix
Okay I am slowly putting the pieces together. I couldn’t find “switch._automatic_watering” anywhere in my config.yaml but it turns out it was in my “customize.yaml” (do people still even use customize.yaml?) so some combination of upgrading hassio and plugging in my wifi sprinkler system must be preventing hassio to boot?
I’ve commented out everything to do with my sprinkler and power cycled but now I have more errors.
Maybe some form of warning when you press “upgrade” in the UI that any incompatible configs with the new version will fail to boot…
Or even better if the config file was checked in the new version’s compiler before the upgrade actually occurs. I know that’s probably easier said than done though.
Okay after a few more things I had to comment out, the web UI is accessible again through DuckDNS but my home-assistant.log file is now 170lines of errors.
All I wanted to do was change from text to icon group tabs. Turns out I was simply using:
“mdi:icon-home” instead of “mdi:home”
There was a bug that also caused this behavior and everyone’s solutions in my searches were “upgrade to the newest version with bugfixes”
Pressing that upgrade button broke my Lightify connection, all my TP-link switches, my sprinkler system (raincloud) needed every entity renamed to remove an underscore, all of my MQTT, my push notifications, my plex and roku connections, and basically every automation I have. And I consider my setup “simple”.
The MQTT “deprecated” is the real kicker because from what I’ve read, the auto-discovery of my ESP32 lighting projects that use code from a project I found that’s no longer maintained or updated is not supported anymore so not only do I have to learn how to integrate the new protocol, I’ve got to climb up on my roof to access and update the ESP32 enclosures for my Christmas lights that have been working flawlessly for 7 months.
Trying really hard not to complain about HA, but pressing that upgrade button REALLY needs to have a warning. I’m on hour 5 of this and just managed to get all my devices accessible again, except a couple I can’t figure out but most automations still need fixing. I’ve wasted most of my day on this and didn’t even get my sprinklers set up
But for the user who follows the guide for setting up HA, copies and pastes a couple examples from docs and a few more from forums, then has no problems for months, one day notices that little “Upgrade” “New Version Available” button and clicks it without thinking…
I’m just saying… a warning built into the UI to check for breaking changes wouldn’t exactly detract from a power users experiences and would significantly benefit simple users like myself.
The real problem though is how simple some of these “breaking changes” are. Half of mine were just syntax or changed config variable names.
I’m sure this is debated constantly by people in much better positions to observe the HA community, but I can’t help but comment about all the hundreds and thousands of examples and projects that people post yaml code for that instantly become useless when different versions use different syntax, most of it seemingly insignificant.
Oh well. I know the typical response on here for complainers is something like “go find something better, or spend hours a day keeping up with changes”
There is a breaking changes addon you can use for this… but… if you don’t read the breaking changes in the release notes that’s on you. This is also a reason for staying reasonable up-to-date so there’s not too much to fix.