How did you install Home Assistant?

Installation
121

Thanks, I didn’t want to critizice your point of view. :slight_smile: Sorry, if that came around that way! :slight_smile:

I’m not very happy with the discussion at the moment, as I find it a little misleading. We as a community discuss this, but from the dev department noone is seen… And I for one would love to get the discussion moved towards a solution. There was an interesting discussion about supporting only one Linux for a supervised installer, not all distros.

In my eyes that would be a good solution, and “the community” should make a vote about that, and depending of the outcome, should stand by that decission and give it to the dev team as a demand, not a suggestion… :wink:

But I know, this is likely not doable. But hey, hope dies last, doesn’t it? :smiley: :smiley: :smiley:

122

300 non-random people, with a specific agenda, just like the ones who voted in this poll. The ultimate proof of the poll’s invalidity will come if and when updater’s stats are published to reveal that more than half of all users are not running Supervised (like this so-called statistically significant poll suggests).

123

Totally agree with @paddy0174 .
All discussion is misleading. We just have to wait to see what the final decision would be. Although I hope to keep the supervised installer but I don’t think they will. To be honest I am disappointed from the “concept” HA right now. The (amateurism) way that of the “statement” of the deprecation it shows a lot.

Even if they keep it, how one can trust that all the time he invests in configuration and accordingly the budget in hardware won’t be fall apart in the next few months for who knows the reason or the logic behind it.

For me this statement was like hearing a “google” voice with nest products which all of the sudden they just shut down the communication with the rest of the systems.

Home automation is in the begging and soon or later will available a lot of systems more user friendly, Instead to discuss this the devs making it more complicated and difficult.

Maybe I am pessimist right now but with steps like these I don’t see HA to be sustainable in the near future. Probably will continue with power-users only, and the rest will choose a less efficient program which can handle.
No problem at all, but I would like someone from the devs to make it clear soon in order for everyone to know where its going.

124

The flowchart is actually very helpful to me. I have been using some version HA for over a year now and still have trouble figuring out what version I am running (core, os, hassio)??? The flow chart definitely cleared things up for me.

1 Like
125

The only reason for placing restrictions on who could vote in a poll would be to influence the outcome of results. By doing that you wouldn’t be getting an accurate representation of the overall user base, just data you want to see. This poll was already heavily skewed towards the most active users anyways since it lasted just three days.

There’s really only one reliable way of gathering this type of data, and that’s collecting analytics from updater. People can opt-out of this so it wouldn’t be 100% perfect either, but far more accurate than a forum poll could ever be. There’d be a super large sample size and it would eliminate any possibility of user error.

2 Likes
126

IMO they should only support/produce a “core” install in a python venv and call it home assistant. anything else should be unofficially supported/produced by someone else. that would solve all this.

1 Like
127

You mean like when I started with ha…

1 Like
128

Hi, New “user” to HA. So far I have tried installing HA Supervisor via docker in unraid. I got HA supervisor up and running in unraid docker and got into HA and looked around. But I could not get HACS with custom components running at all so uninstalled and started looking for other install types.

I am completely new to both docker and any linux style installs however I am not new to computers or coding. Unraid was basically for my NAS but then I noticed the ability to install HA so started looking here.

I must say, as a new person looking in after days and days, I am still scratching my head - I basically could build any system out of spare parts but I am still confused on my options. So I guess as someone else mentioned, if you are seeking the view of a new amateur person looking in - this is bloody confusing.

At the moment I am leaning towards a build (an old surface pro 3 running linux) that I can keep separate from my NAS so I can leave the automation with the house if I sell it, but really I am going backwards the more I read - I thought supervisor was what I should be installing, but now core + ??? or something else. Sheesh.

129

Thanks for a more or less “official” statement. Unfortunately you missed the point in my statement. :slight_smile:

I do have a little experience in managing and coordinating open source projects, and experience shows one thing: if you want to be succesfull in a “community driven” development of software, you should listen to that part of your userbase, that is actively developing your software.

The thing is, HA is relying on input from it’s active user base. Or do you really want to argue with me, that every user with an error has a Github account to report an issue? No.

The people that put time into this project are at most 300 people, I’d say less than hundred. If you loose this user base, you will loose your project. The standard user of HA is not important for the ongoing devlopment of this software - the people that “go the extra mile” and looking and finding solutions, giving support in a forum. That’s the base you must ask.

Let me give you an example: If you take all the 300 users, that voted on this issue, out of the forum, and away from Github, how many issues do you think will get posted on Github? How many questions and threads here in this very forum will get a correct und useful answer? You see where I’m going.

My suggestion is: listen to the people that care for HA, not for all users of HA. Because without the “hard core” group you will likely loose any competition to all other house automation systems, regardless it is FHEM or OpenHab. Only people that have a vital interest in a software and their development are users you can build such decissions (like canceling an installation method without notice) on. If you don’t, you will loose.

I can search for a few projects, that have gone exactly that way, not one is still around. One I remember was turned into a closed source software after a good handful of users and supporters from the forum forked it and left. The last round of investment or risk capital wasn’t very kind to them, they didn’t get any money. The forked open source project is still around, grows and gets better, and has sourced more money than they needed via crowd funding.

And that is not meant offensive against the developer team, it is simply experience. You don’t get anywhere, if you try to work against the vital part of your user base.

2 Likes
130

Remove the base_url completely. It is migrated when you go to 0.110 and is not needed.

1 Like
131

Well I missed the poll, but you can count another for Supervised.
Running on top of ubuntu. Ran on top of linux mint (ubuntu as the base) before this and it seemed to run better. Was going to plan to go back to linux mint, but now…
I think I’m gonna sit tight for a bit.
I do run other/use for other things on this PC and so I really don’t want a HassOS situation.
If I did a VM then I have to spin up multiple and figure out how to get things like GPU passed through to do image processing.
I guess it is all a learning curve and I’ll adapt if I have to, but I’d rather stick with what I have vs. learning how to do it all in another format.

132

This isn’t an “official” poll of any kind.

I just wanted to see how the different install methods are spread over (in this case the active forum) user base. And I wanted the results to be available in a reasonable amount of time too.

Apparently the data will start to be “officially” collected starting in v110 via the updater integration. So at least the devs will have good information on the usage stats to decide which install methods to maintain or not.

It would be nice for them to post those results somewhere on the main home-assistant.io web page too.

1 Like
133

Hopefully a representative sample will upgrade to v0.110.x as well AND not block the stats…

1 Like
134

Gathering accurate information would also require everyone to update to the latest version, which is not the case. A large number of users are 5 or 10 versions behind, or even more. I’ve got a redundant backup running on a Pi that I haven’t updated in months, as an example. The install I maintain for my parents house is always at least a few version behind.

Just in my house at the moment, I have 1x VM and 2x Supervised installs. If I update them all to share the data, it will skew the results as only 1 is the daily driver, the others are for backup or testing.

I’m not suggesting that gathering the data via 0.110.x is not going to be valuable, more that it still won’t be completely accurate.

1 Like
135

I’ve done my part and re-enabled the updater. :slightly_smiling_face:

Though I don’t think it will help keep the Supervised install supported since I don’t use it on two out of three of my running instances. :wink:

1 Like
136

It won’t but it’s more likely to be a representative sample.

137

Still useful to have that data.

138

And which one did you vote for in the poll? AFAICT it only accepts one vote per user, so the results of your own poll are skewed by your own data.

139

I voted for Core in Docker.

And that was per the instructions in the poll that said if you have multiple instances then use the one you consider your “production” instance.

And it didn’t really skew any results since I have two that use core in docker and one that uses the supervised install. So the extras would cancel out anyway.

Besides, it was just meant to be a general information gathering poll, not a definitive peer-reviewed scientific thesis.

1 Like
140

To be clear, I am not a Nabu Casa employee and nothing I post should ever be misconstrued as an “official” statement. I had zero involvement in this particular decision, I’m glad it was temporarily put on hold, and I hope a positive resolution can be found in the end.

My only point was if you’re trying to use statistics to prove the importance of some installation method, it shouldn’t be with a clearly manipulated poll like the one you were proposing. Requiring participants to meet some arbitrary criteria that you decide would just be influencing the outcome to more closely align with your own self-interests.

It’d be better to simply say that active/hardcore users are important to take into consideration regardless of their overall numbers. Which I would agree with.

Sure, I acknowledged it wouldn’t be 100% perfect. I also have a number of instances that I intentionally keep several versions behind for bug testing purposes. And people can disable updater or set reporting to false. The advanced users who like to tinker with every possible setting are probably more likely to disable stuff like that and underrepresent themselves.

I don’t think these factors would significantly impact the results, but we have no way of knowing for sure. Despite these things though, overall it’d give a much clearer picture than anything else.

1 Like