Something is filling my hard drive

I am having disk space issues and was hoping someone could help me determine the cause, or correct me if I am wrong!

Info:
I am running HomeAssistant in Virtual Box. Currently have Home Assistant Core 111.4. Supervisor latest version 228, HassOS 4.10. However I had this same issue with the prior version before upgrading about a week ago too.

Issue:
Originally I set the virtual disk up with 32 GB, but one morning after an unexpected reboot, I had issues getting HomeAssistant started. It would go to the Hassos boot menu showing:
1. Auto Boot
2. Boot System 0
3. Boot System 1
4. Shell

After some digging, I found it might be related to my disk space. Looking at the virtual media manager in virtual box, it showed that I had taken up the entire 32 GB allotted. I upped this Virtual Size to 64 GB and rebooted. This appeared to fix my boot issues, and I haven’t seen the boot menu since.

Problem:
However, I still don’t know what caused the problem. This all happened about 2 weeks ago. I took a look again today and the actual size of my virtual disk is now 37.34 GB. My concern is in a few months I will be back in the same boat with a full virtual disk.

What I’ve tried:

1. I looked through a few threads like these:

Following the recommendations in these other threads:

2. I ran the df -h command shows that the available size is 62.3G but only 8.1G is used:

3. du -hs command shows only 1.7G taken

4. Again from the SSH Terminal Addon I also tried ha su repair. I don’t believe i saw any noticeable change in disk size. Maybe this is because this was run from within Home Assistant?

5. I also pulled the VDI image up in PowerISO to poke around.

The partitions within the VDI image show hassos-data is the largest partition.
306x184

From within that hass-os data partition, it only shows that 10.5GB are taken up:
399x220

  • The docker folder is 7GB (7,516,837,685 bytes)
  • My overlays2 folder within the docker folder is the bulk of that at 7GB total.(7,503,213,066 bytes)
  • The logs folder is 1.8 GB
  • The Supervisor folder is 1.7 GB
  • My database file is only 1.2 GB within the supervisor folder.
    371x460

Current information shown on the VirtualBox Media Manager:

2 Likes

Did you ever find anything more out about this issue?

I seem to be having a similar problem as you, but have no idea what is causing it.

I have similar situation with my Hassio on Raspberry Pi 4 Model B (4GB).
When I mounted one of partition of SD card (sudo mount /dev/sdb8 /mnt) to my exiting Linux system I found something interesting in logs folder. On your screenshot /hassos-data/logs.
This folder contains system logs and the creation timestamps of these files coincided with the graph of increasing disk space usage. I deleted some of the old files.
sd-usage

Unfortunately I have no clue how to clean up these files automatically. Command journalctl not found on my system.

I have not found any additional information. It appears to be slowly increasing still. I’m sitting at 37.50 GB for the actual size.

My logs folder was 1.8 GB back in July. I just pulled it back up and its at 3.9 so maybe this is the culprit. I didn’t pay much attention to the actual size of the Hassio install originally so not sure how close I was to the 32GB virtual size before I upped it to 64 GB. I’ve just been monitoring it every couple weeks for any large increases.

I’m using a Rasp Pi 4B as well and noticed the space filling up on the SD card about a week or two back. It seems to coincide with an upgrade from HassOS 3.X to 4.13.

I took a look and found the systemd journal to be the likely culprit and am watching it now. When I do journalctl -f when logged into the ssh root@<ip> -p 22222 I see a flood of dnsmasq query logs.

I don’t see any easy knob to turn off dns query logging for the dnsmasq add-on though.

EDIT - 2020-09-28:
It seems like the usage has now leveled off around 20% and has stopped growing. I don’t know if that’s just normal log rotation or if me setting debug: false on my zwave: config section made a difference.

I’m wondering how to do this too. I’m looking through my /var/log/syslog file and it’s filled it DNS queries, often the same 5-10 within every second. Seems very dumb to be storing all of that activity, unless debugging. Is there some way to turn down the logging to WARN or something?

Problem solved by comment logger part in /config/configuration.yaml

1 Like

Hi,

I have the same issue. I just noticed that my storage increase ~300Mb at a 2:00 every day… which is exactly when I’m backing my Home Assistant instance using GitHub - sabeechen/hassio-google-drive-backup: Automatically create and sync Hass.io snapshots into Google Drive.

300Mb is exactly my backup size…

I’v checked my config and Home Assistant Google Drive Backup is configured to keep 14 backups in home assistant so disk space should not increase this way…

Anyone else noticed the same behavior ?

I have the same problem my disk is filled from June 10

I would suspect google drive backup. I’ve discovered that when GDB deletes a file on Google, it goes to trashon Google…but isn’t actually deleted for 30 days. I reached my GD limit, even though I had only 4 backups…the rest were in trash and I had to delete them manually.

On the HA side, I would check the log for Google drive backup and check to see what backups you have. Every time you update an integration it creates a backup…if you update three integrations, you’ll find 3 backups.

Poke around in that area…likely you’ll find your problem

I don’t have google drive so what’s problem :frowning:

If you have the supervisor/file editor addon, it will show you the sizes of all the folders. Which folder is growing so large?

I seem to have the same problem? (Or HA is maybe just grown as a OS?)
No backups and I did the:

And:

Minimize home-assistant.log

system_log:
max_entries: 500

So is 19G a normal installation?
I also moved my DB to external MariaDB

I have this same issue, my HA OS is reporting that my entire drive is full, even though I can’t find anything on the drive. I am wondering if HA OS just isn’t managing the disk space well and it thinks its full.

The culprit is usually either database, HA logs, system logs, or backups.

No backups, database is a couple gigs. When I run df from the cli, the overall HAOS files are small, while the NVME drive reports as being very full.

Check your system logs

I feel stupid asking this…where are they? When I get in the CLI via the add on or local monitor/keyboard HA is at most a couple of gigs. Running a df command shows that the NVME drive is almost full though.

I have the same issue on 32 GB space. Can’t find the cause. Did the same diagnostics. Gained some space by “ha su repair”. Occupancy from 90,3% to 81%.

1 Like