My storage levels are RISING. I’m not downloading anything new, but have lost 10 GB over night. I’ve cleared all extra backups and I’ve even cleared out the dbv2 and run both recorder purges [I suspect those do the same thing as deleting dbv2.] I do keep getting notifications that someone is failing to login via the pc I am using at the time that I get the notifications, I don’t know if that is relevant.
Any help is appreciated. I went from 45% of a 128 GB to 65% in less than 24 hours. Also, if I use Balena Etcher to copy my image to a 256 card, will it have access to the full 256 or still just 128? Thanks! Also would a Wyse 3040 Thinclient run a heavy setup better than a pi3b+ or would a 4 be better? Just thinking of ways to resolve this.
Create a new install on the new card. It will expand to use the space available.
Restore your full backup.
You probably don’t need to do this though and it won’t solve the issue for long. 10GB/24hrs is not normal. Do you have nvr or image recognition addons (e.g. blue iris or frigate)?
The easiest way to work out what is eating up all your disk space is to look inside one of your full backups. Use winzip or winrar. Look at the file sizes.
I am running BI NVR, oh snap I may have a snapshot automation now that I think about it. I will look into that. The problem I am running into with Backup and Restore is that I have followed every guide out there to install Bullseye, and I can not complete any of them because they are incomplete and I am pretty sure the community guide one is for anything but a pi. They all assume during random points that you know a non basic thing. I need to run via Debian because if I restore to the HA OS, then I crash every hour and 4 minutes, can’t figure out for the life of me why, whereas on my 128 card with Buster on it I don’t crash. I’ve tried both the 32 and 64 installs too.
If I could figure out how to properly install Bullseye, I’d be flying.
I’ll check the backups
Edit. No snapshot automation. Would it still be doing it anyway?
Why are we guessing? Do you have backups? If so go to the UI and click on the latest backup. Then click on the one before that. What is consuming the most space and what grew the most (in case they aren’t the same thing)? Tell us what that is, that’s where to start.
If you have no backups or the large thing is a folder (/config, /media, /share) then what are the largest files in that folder? Ssh in and run this on the folder
du -ahx {insert big folder} | sort -r | head -5
Share those if it’s not obvious to you what is pumping them out.
I was unable to achieve any sort of expected result using any of the above mentioned methods. Winrar shows me a .json and a tar.gz that when opened just has two more .json. So no files or folders to see the size of. Looking at the backups in the UI does not break down the size of individual files and folders and upon SSH in to see if it was /media, this is what I received.
du -ahx /media | sort -r | head -5
4.0K /media/usb7
4.0K /media/usb6
4.0K /media/usb5
4.0K /media/usb4
4.0K /media/usb3
entering via samba shows every folder as empty except /config and /backup.
Treesize shows the same.
not really sure where to go from here, but it rose another 5.5%. Found a snapshot folder and deleted it, but nothing had been modified in it in weeks.
Oh right, my bad. I do all my backup management in the Google drive backup addon and that shows the size breakdown. I forgot the native HA UI doesn’t show that.
That’s weird. Do you encrypt your backups? If so make an unencrypted one for investigation, the encrypted ones can’t be expanded as easily outside of HA. The unencrypted ones should just be normal tar files though.
Well that’s weird. How about your logs then? Go to settings → system → logs and see if anything is spamming something rapid fire. Don’t just look at core logs, click the drop-down in the top right and check out each one. See if any are really spammy.
So I went ahead and checked my Google Drive backups as well. Nothing shows to be large, but folders don’t display their size either. All Unencrypted.
I checked through the logs, nothing particularly spammy. Checked every section. Half were empty.
I did the same dirty install of supervisor onto Raspian Bullseye and moved it onto the 256GB. This time I left EmulationStation and DIYHue behind, as I no longer use them.
I restored the backup - 235GB available.
39 Minutes of Uptime - 214.8GB available.
I have noticed that if I completely reinstall and restore a newer backup, made AFTER the disk usage has deminished, the storage always returns to 235GB available, but then rapidly drops.
IT GETS WEIRDER!
This whole thing has caused me to start experimenting with other installs. I pulled my old desktop out of the closet, factory reset it, threw VirtualBox on it and installed the OS the official way, cloned the VGI or whatever format it is, raised the storage space for the VM to 256GB, restored my backup and 235GB available… I expected this. I did not however expect it to have not changed more than 0.1GB running simultaneously as the Pi, both started at the same time with fresh install and same backup.
I’m completely confused.
I ordered a RPi 400. I’m hoping the reason it is working on the VM is because of the type of install. I’m hoping that the reason the Pi3B+ was crashing every 1 hour and 4 minutes when installed on the HA OS was because of the RAM or the strength of the 3B+ and that HA OS will stay stable on the 400. Fingers crossed. I hope I didn’t just throw away a restocking fee.
Will update as things progress.
Also, should I just cancel my order and leave it running on the VM?
Any idea how much difference in power? Or should I run a energy outlet on both to see if it’s big enough difference?
I was gonna do the supported OS on the 400.
Well, I am pretty sure I figured it out. Will monitor and mark as solution if correct. I swapped the IP of the VM and the Pi to make MQTT run though the VM. I am running ESPresense and it is receiving more MQTT notifications than ever. I can probably remedy this with the filtering feature in ESPresense, but am heavily considering switch to Room Assistant.
In the event that I stay with ESPresense, anyone know how to get it to clear whatever is filling up?
Edit: ESPresense fits my needs better than Room-Assistant. RA can’t read the signal from my animal’s Tile trackers AND be connected to the Tile App. ESP allows for both. I finally got most of it calibrated. I just need to figure out what data is being written due to the MQTT from it and how to purge that? recorder purges did not reclaim the space.