Looks like the GitHub issue was locked so responding here.
I sympathize with the HA devs as a developer myself. Bugs happen, it’s just a fact. My argument is that certain scenarios should probably be tested a little more thoroughly. This isn’t some obscure scenario AFAIK, it affects anyone running HA OS (or whatever it’s called now) with a Pi 3b+.
I do have an issue about the tone many are taking regarding this issue. Seems like some devs and mods are taking a very hostile stance here and I’m not quite sure why. As far as I can see, that vast majority of affected users are just trying to report what they’re seeing and getting responses like “you’re not describing the problem correctly” (boot loop vs core crashing, because the average user differenciates the two of course), or “maybe you should be a beta tester”. It’s not really constructive.
Regarding becoming a beta tester, I actually wouldn’t mind doing that if it helped the community. I know most people saying this are just using it as a device to victim-blame, but I’ll take you all up on that. Is there a guide somewhere where I can do that? I understand the mechanics of opting in, but how would one run their production (stable) instance in parallel with the test instance? In my case I have a z-wave stick in my Pi 3b+, but beyond that nothing too crazy. And not just running in parallel but also how to properly stage and mirror changes between the instances. I’m genuinely interested in helping out here if given some reasonable direction.
Because we’ve been urging people to join the beta to find these things and no one does but the same 20 people. Then we come here and there’s many people saying “Why aren’t they testing more”. Who’s they? I’m doing my part to help this open source community by running the daily build (the alpha) and I mod in my free time. What are you doing to help? What are others doing to help? I don’t recognize anyone who is complaining about this, so my assumption is: You’re doing nothing but complaining. And that annoys me. I wager many other people in my position feel the same way. Remember, this isn’t a mega corp, this is an open source software with severely limited resources.
This is the tone I’m referring to. As I mentioned, I’m more than happy to become a beta tester just tell me how I can help. I’m legitimately asking how I personally can help contribute, but it seems that everyone would rather blame me and others than actually guiding those willing to help address the issue.
EDIT: the “they” I’m referring to is the core HA devs. The ones who maintain HA and are responsible for testing and release. The same ones who made the decision to pull the release. “The people in charge”. As a mod, you may not be part of that set of people, but you are responsible for proper moderation and victim-blaming those affected by a system-breaking bug (especially one legitimately trying to take you up on your offer to beta test) isn’t a great way to interact with the community.
Much appreciated. I’m hoping the instructions include how to run the stable instance in parallel (on a separate device is fine if required). Opting into the beta on a production instance is a non-starter and if that’s the case (again, hope it’s not), that’s probably why there are so few beta testers. If it’s pretty painless to stand up a test env to mirror my production one, I’ll absolutely help beta test and push others to as well however I can.
As mentioned, I understand the mechanics of upgrading my production instance to the beta channel. But the beta channel isn’t stable enough to run in my production instance. So for me, and I expect most everyone with a desire to help beta test, I need a way to run a production (stable) instance alongside the beta instance with more or less mirrored configuration. A choice between having a stable working instance vs helping beta test and putting my system (and thus home’s security) at risk is an easy one to make. I need to be able to have both.
The link that @dshokouhi posted has the directions for both the alpha (development) and beta.
You’d have to run it on a second machine if you don’t want it on your production machine. I personally run the alpha on my production machine and I rarely run into problems. I think I’ve ran into 1 issue that caused my frontend to to fail. Zero major issues otherwise. I reported the frontend issue and downgraded to the previous alpha and was up and I’m running in minutes.
The beta is only available the Wednesday before release. The alpha is built every night. I commonly refer to it as the “nightly” or the “daily”.
Yea running it on a separate machine, especially the HA OS install path, seems completely reasonable and required. Running the nightly on the production instance probably wouldn’t pass the Wife Acceptance Policy though .
I did some thinking and I think most my integrations are easily multi-instance without issues, but I’m not really sure how to handle z-wave. If I were running z-wave js on a separate machine or in a way that was exposed on my local network (my understanding is that the addon isn’t accessible off box; would love to be wrong though), then I could point the integration to that shared place but seeing as I’m not currently, and rearchitecting my system seems to defeat the purpose of trying to test this scenario, I’m a bit stuck.
At this point this discussion might be better moved to discord or something though.
Well this exact issue would have been hit but earlier for me, which although good for the community would still have caused just as much disruption in my house.
I guess it doesn’t make sense to me that the solution to help avoid bug is to opt into less tested code
Well it would have been caught. You would then downgrade to the current release. I think you’re afraid of the word beta, the core of HA doesn’t change much. Breaking situations like this are extremely rare and fixing the problem is literally 1 command line call.
.