New build…?

Hey all,

So I’ve had HA running for a few years now and it starting to get very messy with all the different things I’ve tried over the past few years (some successful, some not so successful).

2 questions:

I am considering wiping my Pi and starting from scratch. Any advice from the masses on undertaking a slash and burn?

I would like to upgrade my pi but don’t have the cash so do so right now… would starting fresh again now, then moving it over to a new pi be doable, or should I just wait until I get a new machine?

Thanks!

Absolutely doable, just use the backup feature.

1 Like

Cool thanks mate, I suspected as much, but have never had to do it before. Means I can get started on the rebuild now!

Why move from pi to pi?
Why do you feel move is needed?

There are just so many options but really your end goal determines what is best. Is another Pi the best you can do for same cost or slightly more? Just be careful not to move to another temporary home.

1 Like

Was considering getting one of the HA green or yellow… the pi I’m currently using was already pretty old and has been going solid for the last 3 years running HA. It’s starting to lock up so I have a smart switch on a different system, to reboot it… etc…

Will do a bit more research on the replacement machine when it’s time to upgrade. Really I need to make it a bit more stable and easy to use, as I don’t have as much time to put into fixing bugs and maintaining the system as I used to.

Actually, now that I’ve started digging into this I’m getting stuck on the best way to keep some of the more complex integrations for the new system. For example, the BroadLink UniRemote Integration. I have set up a lot of learnt inputs, scripts and automations. Is there a simple way to pull out the BroadLink integration and save it to be incorporated into the new build?

The config folder stores all of HA’s files. Any settings, passwords, for all integrations are here.

A HA backup is just a zipped copy of the config folder really.

If you have any external service, that will be stored separately wherever you placed it. External database, external mqtt, and similar. You would be responsible for making sure that data moves to new system

If you have customized the OS then this will need to be considered. WiFi setting, externally mounted hardware, and the like will need to be setup with new install.

These basically hold true for every install method. OS, docker, vm, etc.

Generally keep old system running until new is setup.

Restore for me is like this.

Setup new hardware with default HA running. Verify I can make HA boots and runs with no errors. Install the same version currently running on old hardware. You don’t want to add problem by installing version with breaking change.

Add any OS customizations. I have udev rules for zwave stick and coral to mount with specific location and name. I have some drives I need fstab rules to mount. I take this time to setup temp IP for new hardware(I may keep this IP or later when HA has basics running I may move to the standard HA IP I have as default in my network). Each of these items is verified before move on.

Next I stop HA (both on old hardware and new)and restore the config. I use docker so this is easiest way but if using backup you can keep HA running and restore. It is at this time that i move zwave stick, coral and other physical device to new hardware. These items need reboot to be recognized so this is good time, also new system is ready and old can be stopped. Presuming you know what your doing all will go well but I plan at least 1hr downtime to allow troubleshooting short period. You can go back to old hardware if troubleshooting goes over 30minutes or stay down and press forward. When HA is back running I just starting checking what is or isn’t working and resolve errors.

I use external DB. With the config restored it should become active. With the exception of integrations that use a physically connected device (like zwave) you can mostly run two HA instances side by side but NOT using same IP.

Check log if problems occur.
Always good to verify from HA that you can ping device if not connecting to it or access a folder if things are not working. Depending on install method incorrect file permissions will cause errors until you set the correct permission. This is actual purpose of booting to default before restore. It gives sanity check so you know it can run and allow you to verify expected group:user and location of files. If you don’t do full restore often this is useful hint.

I get where you are coming from, I’ve considered starting over myself. I figure the first year was spent really learning things and creating an awful lot of “stuff” and now that I’m more seasoned with HA I can approach my home automation in a better way.

It’s understandable why you stay on the rPi, if it works then it works. I outgrew mine but if (or when) I rebuild my HA it’ll be on the same system it runs on now with a full unzipped backup of everything I had before so I can copy/paste things that work well without re-creating the wheel all over again for every script and automation (and, and, and… lol).

I figure I’ll start on the project, then restore the old system and just keep pecking away at the new system through backups and restores until I’m ready to make the new system the only one. I have a sandbox HA but since my Zigbee and Z-Wave need to be on the main system I need to backup/restore or I don’t really get the full picture - although I run on a VM so perhaps this wouldn’t be an issue at all.

Thanks for the detailed response mate… it sounds like your setup is a lot more detailed than my own. I’ll see how it goes!

Thanks mate, I think on I’m a similar journey as you.

I’ll look at pulling that backup so I can dig through it.