Significant improvements to backup & restore functionality are coming in the 2024.2 release.
8 Likes
The release is coming today. I’m curious to hear from anyone, once they installed all updates (core + supervisor), if their experience was better.
Had a quick look and can’t see anything mention in the notes for the release - what’s new?
I’ve since switched to a mini pc and proxmox so I can do snapshots and backups on top of the HA built in backups and it’s very much more reassuring (a few seconds to jump back) to be honest.
I’m comfortable of having different levels of backups (after all I restored my PI4 config to the new miniPC and it worked well) but the original request is valid - some sort of log or indication of how it’s going would be super welcome.
Majority of the changes happened in the frontend or supervisor but there were some in core too. These aren’t mentioned for some reason in the release notes.
1 Like
Edit - read the posts before mine, going to wait and see if this was addressed.
Voted. No improvement for me – this is the first time i’ve had to restore (i was on the latest 2.3 i think it was) and there is NO indicator if the backup is ongoing or not, if it’s done, what’s going on. I’m just sitting here wasting my time staring at a screen.
1 Like
Can confirm. Just bought a Pi 5 and did a 3.5GB restore in maybe 10-15 minutes (amazing!) but did not receive any indication that the backup was completed.
1 Like
Progress indicator, or better, some form of basic log output so we know it’s doing something and not stalled.
Support for states are actually there in the Supervisor API but no one built the Frontend elements required for that yet.
I agree something is needed to let you know where the restore of the backup is. A while back I got burned by an update (or so I am led to believe) and the restores which for me take 10-15 mins went on for close to an hour. I had enough and re-built and had another backup which restored most of my stuff.
This is needed. Doubt my half cent will help the cause but just adding my own experience here.
1 Like
Could it be that backup encryption is an influencing factor?
My last recovery of a backup on an RPi4 took more than 2 hours, and the backup is only 340MB.
I then setup a VM on my Proxmox server and recovered the same backup, took still nearly one hour.
I setup another VM on Proxmox and recovered a backup from two weeks before with 300MB in size, and that was done in 20 minutes.
The only difference between those two backups besides the 40MB in size were that the smaller one did not have any password on it, the other one had.
I have now also voted for the progress indicator, watching 2 hours at “connection lost, re-establishing” was a bigt frightening.
15 minutes of backup recovery for 3,5GB on a RPi5 do not sound right to me, not sure how that should work in relation.
I’m currently restoring my backup of ~500mb on a Raspberry Pi 4 and I still do not have an indicator of the progress. I’ve had it run earlier and I left my workspace to go do things. I came back and it sort of didn’t continue. So I’m trying it again and I still don’t have anything to indicate progress.
I get that an actual progress bar would be very hard to code seeing as it might be the database or an addon or whatever and you cannot take into account everything.
A way to give you an update would be to show whats happening in view only console window. Or a simple steps view like 2N does for example (linux based):
My restore is running. Found out that if you have a monitor and keyboard attached, you can use the cli to do: jobs info
It gives you running jobs. If you see an empty list, it is not doing anything. If it is busy, you see what jobs it finished and which are running (such as restore addons)
1 Like
I have tried to restore my HA instance several times, since I cannot upgrade from 2024.4.3. However, I have successfully upgraded to OS 12.4.
My backup is 5 GB.
Ryzen 1600X with Virtual Box on Ubuntu 22.04. In Virtual Box, I have set 6 cores and 8 GB RAM (total RAM is 32 GB).
Any help is appreciated.
I did somehow manage to restore my backup of 5 GB. However, it took several hundred of GiB:s flowing to HA. What HA did with all of the GiB:s is unknown to me.
It took several hours for me to get to the screen where I could choose to restore everything or choose what to restore. This time, I choose to restore everything and after that it did not take that long before I had my backup up and running.
Host: Ubuntu 22.04, Virtual Box 7.18. Ryzen 1600X with 32 GB RAM.
I do not know exactly what did it, but now I can restart HA without having to kill VirtualBox. I guess some of these settings did the trick in VirtualBox.
After a couple of hours trying the new HA out, it seems that I have got the same problem. I cannot restart HA without killing the process on my host. I guess it must be something with VirtualBox. All I tried to do was to add InfluxDB, which did not restore correctly and I had to supply it with a database name/password.
Well, now I know a bit more than I did a couple of days ago. I will try KVM instead.
I’m not sure you will get many answers in this section as it’s for feature requests rather than technical assistance. Maybe @moderators can move it for you ?
2 Likes
Not really the indicated feature but a way you can know if the restore process is alive, is by opening up the browser console and heading to the network tag, the backup process seem to ping the browser every certain amount of seconds, not perfect but at least it does let you know something is happening.
I totally agree on this.
I need to restore a 3g backup file on a 4core i7 but I have no idea if its actually doing it or not - i just get a spinning wheel
Truly dont want to brick my machine by rebooting it. The Observer is still up and running.
The Observer would be the best solution for it I agree
Having UI output and/or a log of the restore process could also enable optimizing the whole process. I can’t help but think why restoring a 800MB backup on a pi4 should take multiple hours.