Upgrade within HASS.IO from .89 to .101 blew up instance

I normally back up all of my settings and all, which I did in a snapshot.

Sad to say, the snapshot will not download and sends my RPI3B off into not responding land.

I was able to download/backup the previous snapshot from last month with no issues after I power cycled the unit.

Now, I click on tot he upgrade button in the web interface and all contact with the unit goes away. I waited 4 hours and discovered that the unit was hung. Again, Power cycled it and it came up as .101. Great! Went in to double check some things and it rebooted on me. When it came back, it auto-rolled back to .89
Huh ??!!

now after attempting to do another upgrade to .101 the unit answers a ping but no interface and no SSH (it was installed).

Basically, .101 killed my RPIB :frowning:

Does ANYONE know of a way to re-restore this ??

thanks

There were a WORLD of changes between .89 and .101. It’s not surprising that it blew up. What you really need to do is look at the log file, find out what is broken, and repair the configuration one piece at a time.

Thanks.

might be easier to install .101 from scratch and just reconfigure.

Not sure if the snapshots are even viable.

Went to the console of the Raspberry Pi 3 B and was able to do a snapshot restore. Still no SSH no interface… nothing.

Any tips for forcing a clean re-install from a snapshot ?

thanks

Snapshots don’t provide you with a way to start over. You would just reflash your SDCard and start from scratch. Snapshots are intended to restore to a point in time.

thanks. so like only within the web interface can you wipe and restore cleanly ?

If I reflash, can the config/snapshot be used to reload ? I woudl really hate to lose all the settings, interfaces, modules, etc. and have to start from day 1 again. yuk !

You should be able to restore a snaposhot from an old version but… you have a LOT of config errors to fix I’d say and thus using the log to work out what is wrong is probably the best approach.

You’re conflicting with yourself here.

Do you want to have your setup or do you want to start fresh from scratch?

I feel your pain. Before I kinda figured out what I was doing and got into the habit of a robust backup procedure in place I had to rebuild my system from scratch not once, but several times. Lost weeks of my life. Once you get back up and running I’d recommend you do one or all of the following ( I do them all).

  1. Start using a snapshot backup addon. You’re screwed anyway if your SD card craps out and your only snapshots were on it. There are options available for both Dropbox and Google Drive. I switched to the Google Drive one available on HACS.

  2. After I’ve made any changes to my configuration, I copy the entire config directory (except things like the db and tts cache over to a folder on my Google Drive. If I’m doing a lot of work this gets done almost daily. I use samba so it only takes a second to drag and drop the files. You could probably actually automate this with a shell script if you really wanted to.

  3. Use github. I’ve only recently started doing this also is a good way to keep most of your configuration backed up. Remember any stuff in your .gitignore files won’t be there though which is why I still mostly rely on #2.

BTW - I agree with @DavidFW1960. Start picking apart the logfile and have a look at the release notes to find all the breaking changes for hints and what needs to be done.

Good luck!

OK, an update.

I could not get SSH to run on the .89 that was saying it was .101, but that’s ok. I had a backup of the snapshot off on another system (backupSys).
Thankfully the snapshot can be unzipped/untarred, etc. and read on a local HD !

I took the SD card and reflashed it. It now runs 0.101.1 with no configs… :frowning:
But I did not despair !

I reloaded the add on for Samba and for SSH and I could connect !!

I then copied the /config files from the snapshot (yes after backing up the origs) to the new instance.

Rebooted an waited. Some groups did not transfer well So I need to set them back up.

but the automation, the switches and everything else (minus the environmental probes) all came back.

For some strange reason, had to rename the main switches, but I can live with this.

I’ll find out later tonight if the automation works as it used to.

Thank you ALL for your quick responses!!

YES, I even copied the DBs over. Seems that if I didn’t the switches and everything else just would not show up. bad/good idea ?

thanks again!!

thanks for the insight. I responded to myself on what I did. Maybe create a doc to restore easier ?

cheers!

You don’t need the database. It makes no difference to anything showing up. It just has history in it and generally restoring that will give you lots of grief with errors in the log so you may want to delete it.

Your switches might have been renamed due to some changes since 0.89. You will likely find stuff not working in automations if you don’t work through the release notes since 0.89

Thank you for the response.

Just now trying to figure out WHY none of the automations work :frowning:

set up tests for time and nothing. strange.

Got the Automations to work…

In the latest version (0.101.1) that I reflashed, the Actions changed from “Call Service” to “Device”. This is where the device is told to turn on or off and not a service.

Although, one would think that the service “Homeassistant.on” for a device would still work. oh well another thing to make you go … hmmmmm

thanks to all again.

one would think that the service “Homeassistant.on” for a device would still work

This works just fine with any entity.

- service: homeassistant.turn_on
  entity_id: input_boolean.my_boolean

- service: homeassistant.turn_off
  entity_id: input_boolean.my_boolean