I have a HA currently running on a 32 bit kernel and wanted to upgrade to 64. Therefore I did a complete backup and downloaded this (this setupup used a supervised installation method done by a friend - just in case thats important).
Now I wanted to run it on an rpi4, so I chose the home assistant os in the official raspberry pi imager and booted it up. I was greated with a onboarding page and uploaded that 8gb backup file. After choosing restore I went away only to come back to an machine not responding.
So I started over and only restored the core 2023.4.6 and it took more than 3 hours, after which I decided to abort and ask for help here. Is that normal to take so long? Is it compatible to use a supervised backup on an haos system or do I have to use supervised again? Is that possible on RPi, itās not listed on the install pageā¦
Thanks a lot for clarification.
Is it possible to mount a nfs folder directly to the media location? At the moment I have my nas mounted to the folder frigate nvr uses to store itās files.
Well I made it to the cli. Somehow the WebUI got unresponsiv.
In the cli I can use backups to show available backups and it shows only the one I uploaded with about 8gb size, my installed addons and name āUmzugā.
How do I restore from there? backups restore Umzug gives Error: Backups does not exist
In the help example it uses some hex-names to identify a backup, but I donāt see such identifier.
ok I found it - slug
When starting the restore with this it finishes way too fast (<5s) and I donāt get an UI. Do I need to set a flag for complete restore?
My full backup is 2.4GB and I have waited 2 hours now. WebUI is still unaccessible but the observator URL http://192.168.1.201:4357/ is accessible.
Sometimes it shows connected/supported/healthy but sometimes it shows disconnected and a lot of logs from the past? The log stopped midway seems itās too long to display properly. This is a fresh install and i am restoring from the full backup I made earlier today.
With a monitor attached, there doesnāt seem to be any signals. I just power cycled the machine and it is still the freshly installed version. Restoring didnāt happen at all.
Well I can confirm that the monitor does not show anything during the restore. The webUI just goes offline. In my case it took more than 3 hours until it returned with a completed restore. So my problem was just my impatience and a lack of progress indicator.
Thank you for the info. I will find another day to wait 3+ hours then. I also just found out my backup has different version of OS and Core than the latest HA OS install. I should have made sure my current HA is updated to the latest before I proceed.
Can confirm that all you need to do is have patience and wait. Make sure nothingās locked up using the observer URL, but other than that, just let it do itās thing. It took me about 1-2 hours with a 800mb snapshot on an RPi4. I imagine this time would increase quite a bit for larger backups.
ā me when itās 1 am and I canāt get any lights to turn off. My SSD jumped ship and now I am trying to restore my backup. But itās taking for ever.
Is anything new in this regard? My 400MB (only HA, addons, etcā¦without database) file took good 2 hours to restore, and itās Intel NUC with i7 cpu with 32GB RAMā¦ it should go darn fast in such powerfull machine (HA restart is under 1min).
I browsed a bit on linux via console while restore was working, and it seems that copying from āunknown locationā to final folders is teeeeribly slowā¦ once docker folder increased a few kilobytes per second. So, i guess that something is ābrokenā inside HAOS or HAā¦?
I also did a restore which I decided to abort. First after around two hours, then after reading here after around 4 hours.
Iāve then decided to make a firmware update of my RPi4, as home assistant was offline anyway, as I typically do such firmware updates and as I have a NVMe over USB connected to the RPi4.
After this firmware update, the restore was done already the first time Iāve checked the status, which was around 1-2 hours after starting the restore.
I hesitated to post this as I didnāt make a precise time measurement and as it doesnāt explain to me why e.g. also on other hardware the recovery process takes that long as e.g. here described for Intel NUCs.
But I thought Iāll post it as Iām quite sure that the recovery was much faster in my case after this firmware update. Iāve updated to the latest firmware version and I had a version from 2022 before on my RPi4, but I canāt tell the exact version any more. I didnāt expect such a dependency, otherwise I would have noted this information.
Do you know how do I check the current firmware?
Iām running ha os on an SSD
Edit:
I was able to check the version and update.
I was running a very old version.
It could be a placebo, but at least for the backup it looks like there was an improvement.
Yeah. I have same issues on bare metal (BeeLink Mini PC) and on KVM on a NUC i7. On the NUC it takes about 45 minutes. Nothing special about the backup (300mb). On the VMs even longer and they are just bare-bones dev/test instances (90mb max).
I had to learn patience but there is a problem with the restore function.
ā me when itās 1 am and I canāt get any lights to turn off. My SSD jumped ship and now I am trying to restore my backup. But itās taking for ever.
