Given that we are never getting yaml back - what is the recommended way to “rebuild” a partially broken system?
I know I’ve asked this before - but rather than have an informed conversation, the topic just gets closed.
Historically, when the configuration of Home Assistant got messed up, for example, a heap of failed devices that have been replaced - I would just delete the relevant .storage files, restart, and everything would be loaded from yaml.
This can’t be done any more.
As far as I can tell - we are expected to “start from scratch”. Using a painfully slow GUI to re-add every single device again.
Am I missing something here? Is there a “dump devices but retain configuration” tool or command somewhere?
What is the “Home Assistant” recommended way to resolve this issue?
There’s no need to touch .storage files at all if a device fails and needs to be replaced. Just use the GUI to remove the entity / device and then add your new / replacement device.
Playing with .storage files is why you are having trouble in the first place.
The problem I have is it can be tens or even hundreds of devices.
I’m not going to go through and do a replace - and for platforms like vera that’s not even an option.
I’m still lost as to why we’ve gone toward something that makes it nearly impossible to easily “start again” - in the name of presenting everything in a gui.
Well, there’s a few reasons - dynalite and vera have both caused me problems before
I guess more importantly - when there’s a simple, no fuss way to “start again” why not replace it if something is getting deprecated.
The question still remains, how can I drop all the devices, but keep all the configs?
I understand your itch. I have similar issues with the rfxtrx integration and auto discovery. It’s a pain to remove devices and deleting the integration requires me to redo all my devices and entities.
There are so many “weird” errors in HA - sometimes - just dumping and starting again is a totally valid course of action.
My problem now is - I can’t.
It literally takes ¾ a day to re-add everything via the gui - when it used to take 0 seconds by having everything in yaml
I’m just lost as to why key functionality (especially for those that develop things for HA) would be dumped - when there’s a far better way to do it - keep yaml - but have the UI update the yaml instead. (for those that don’t want to edit)
Look I intend to agree, but I think you are pissing in the wind, and uphill too
Yeah Tinkerer is a massive non-fan of mine - doesn’t like dissenting conversations
I just don’t think much thought was put into the whole deprecation of yaml. I definitely don’t think the conversation has been “done to death”. In an effort to move everything to the GUI, we’ve lost some pretty important functionality.
I think healthy disagreement is fine. I think that once it’s clear the developers have moved on from a decision and that they’re not interested in revisiting it, then at that point it’s pretty silly to keep trying to raise an army with pitchforks…
Whether or not that first statement is true is irrelevant. The devs have made their position clear. If anybody doesn’t like it they’re free to fork HA and do something about it.
The topic however is done. Dead. Pushing up the daisies.
Instead of complaining about how you disagree with the developers, why don’t you do something about solving the problem you’ve identified?
Sure, that’s what moderators do. When things become repetitive, hostile, or just starting becoming nothing but a group of people complaining we discuss the thread and then one of us locks it.
So, don’t turn this into yet another whine about the deprecation of YAML and instead do something productive.
I’ve had very hit-and-miss luck with snapshots. Additionally, this is less about a “healthy” state, and more about there has been some major-ish change.
A lot of the time I find something I work on in a dev environment that works fine - ends up in one of my HA instances, and then can mess things up (for example).
I guess - yes - but again - that’s a massive pain.
If I’m to replicate even part of a production environment - that’s hours spent in the UI adding integrations.
I’ve historically been able to start a clean dev environment - and with the copying of a few files, replicate (near enough) my production environments.
That’s what I am trying to achieve here - somehow replicating what’s lost.