Is there a notification when the upgrade is done or how did you discern that it is?
can we somehow move from yaml to UI and keep the collected data, in this case for the utility_meter entities?
I wouldnât like to kill my longer term counters
utility_meter:
watermeter_monthly:
# name: Verbruik per maand
source: sensor.watermeter_stand
cycle: monthly
watermeter_quarterly:
# name: Verbruik per kwartaal
source: sensor.watermeter_stand
cycle: quarterly
watermeter_yearly:
# name: Verbruik per jaar
source: sensor.watermeter_stand
cycle: yearly
edit
yes we can:
as posted on Discord #integrations:
I commented the yaml and created new ones with the exact setting/naming, but added a suffix to the entity_id (note the names are Not equal to the entity_id, that was my main challenge). During startup I was able to edit the entity_idâs and ditch the _ui suffix. All entities survived and have their collected data. Nice
2 Likes
You are right. Leading edge is bleeding edge 
1 Like
Same here, i had to go back to last version backup 
1 Like
Yeah, it was my bad. Got the google calendars integration in my custom components and after deleting it everything works fine again. Thank you!
I didnât see any notification to notify me the migration was done. So I guess you have to try from time to time to restart HA until HA let you do it.
1 Like
Iâm using HA as a docker container, do you know whether it also prevents me from rebooting in my setup?
If the database upgrade is still in progress pressing the Restart will fail, but you could just restart the container, not sure how your database would hold up tho.
Hmmm, Iâm a little scared:
I have quite a big database, but it seems like it lets me restart. Started the upgrade about an hour ago.
Edit: Seems like itâs hanging (2022-04-07 17:27:14 ERROR (MainThread) [homeassistant.components.recorder] The recorder queue reached the maximum size of 30000; Events are no longer being recorded)
What database type are you using?
The Samba addon was up and running. I believe I checked the logs and they were fine. I wonder whether a process can lock access to a folder while something is happening, such as a DB update, as my ability to browse the folders (config and backups) seemed to work and not work randomly.
Also, I was getting errors that might be explained by the system not having access to configuration.yaml as it complained about default_config: missing, recorder, logger, etc not being setup etc.
MariaDB in a docker container. Iâm currently trying to figure out whether the upgrade has already been completed (by looking at the current schema inside maria itself).
Edit: It doesnât seem to be done yet, also doesnât seem to do much right now:
(attributes_id column still missing in the states table)
Awesome release! Thanks!
Just quick question, is there no way (yet) to edit a UI group after creating it? Maybe I am just missing it, but I canât seem to find an option to add or remove an entity from a UI group.
Give Maria time to chew on the command then.
My database was around 6GB before this release it didnât take too long to add the column.
I would love to join the beta and report issues but I have run into considerable issues recovering from previous failed updates, and have found that restoring from backups does NOT always restore you to a functional state. With this latest update, I tried to restore 2 or 3 different backups and things were only getting worse.
I really need to educate myself on taking a backup of my VM to ensure I can actually fully downgrade. I also need to learn how to save what I need to recover without major work such as my dashboards, configurations and Node Red flows without relying on HA backups as a secondary recovery mechanism because it would take me months to re-develop what I have today.
My knowledge is not enough to understand what is failing, and above all, how to recover from it, so I am very apprehensive of testing betas on my production system even though I would love to contribute.
I have a secondary test system but for now it has not been of much testing benefit as it has no radios on it, and even if I added them, it would not have devices connected to them. I updated it to 2022.4 betas without issues, yet my production system had serious issues (mostly ZHA on its knees).
I suspect that in some cases I may have added to the issues by not allowing background processes to complete. I wish there were more visibility in what is happening during an upgrade. Someone here recommended adding homeassistant.components.recorder.migration: debug to the logger config which I did. Hopefully that will help in the future.
I donât want to get into the specific case, but I do feel that some comments are quite aggressive. I have been using HA since August 2021 and have tons to learn. I may ask dumb questions at times, or seek help on stuff that is documented somewhere I am unaware of. I sometimes likely donât even know what to search for, etc⌠you get the idea. Being scolded on the forum by those who clearly know more than I do and are annoyed by what they consider dumb questions is not the best way to build a helpful community.
Right now I am on the receiving end of help, but I intend to be on the other end as my knowledge grows but I will always be respectful no matter how trivial the question may be. I also understand the frustration that people go through when dealing with issues, as I keep getting a good dose of it too. I love the project, and enjoy the frequent updates even though things sometimes break. In the 7 years with Smartthings, stuff did not break as often but progress was at a snailâs pace. Definitely not the case here 
Loving it!
10 Likes
Yeah I will wait. Although over two hours for 5GB of data sounds weird. Also the maria process isnât showing any significant CPU usage.
Edit: It seems like the NULL process ID 14 is actually in a transaction causin the lock (?)
By connecting to the docker containerâs mysql listening port using the mysql binary. (Something like mysql -u USER -p -h 127.0.0.1 -P 3306)
1 Like
I am running HAOS in a VM so I am guessing I canât do that. The only console access (Terminal) I have is the one specific to HA which appears to only take HA specific commands.
couldnât say really, maybe give Mariaâs container a restart