Disk space goes drastically low

Hello, today after updates my diskspace goes low very fast even after I delete 60GB of files.I have an SSD with 240GB volume and disk usage utilities shows it used only 105GB of files total but free space is about 6GB now only. There are many unaccessible folders for disk utilities so, which should I purge?
Also my /proc folder shows 128 TB of space usage.

I already tried to clean all possible with bleachbit

P.S. today also I installed Android Studio and Android SDK. Also Gradle downloaded some gigabytes of files but it’s all is about 20GB only

People said /proc is not needed to clean

Personally, I would not use “bleachbit”.

You can ignore that 128TB for “/proc”. That’s virtual space, not real space.

You very likely are using “btrfs” for your file system, since that’s the default. It is the nature of “btrfs” to use space.

As an example, I just checked Leap 15.1 that I have running in a VM, with “btrfs”. The “df” command shows 15G is used. But when I do:

du -s /*

to get the totals for the main directories – and then add those up, excluding “/proc” and a few other “tmpfs” file systems, then I see only around 8G is use.

The additional space is tied up in snapshots. Keeping snapshots is what allows a rollback to an earlier version. But sometimes these can get out of control.

With Leap 15.1, you are probably updating often, and this leaves older snapshots with lots of older versions of the updated software.

You can use Yast Filesystem Snapshots to list the snapshots. And then you can delete some to free up space.

Personally, I don’t have much experience doing this. I have mainly stayed with “ext4” for file systems. My virtual machine with “btrfs” is only an experiment and learning experience, but I will probably continue to use “ext4” on real machines.

To get proper info on btrfs, use


btrfs fi du -h /

and


btrfs fi df -h /

Last time I remember my Snapper works correctly was on Leap 42. Now it has a kind of “The config ‘root’ does not exist. Likely snapper is not configured.” error since my clean installation of Leap 15.

/usr/sbin/yast snapshots
No such client module snapshots

If i do it right.
And looks like I need to clean something

/usr/sbin/btrfs fi df -h /            
Data, single: total=221.54GiB, used=216.45GiB
System, single: total=32.00MiB, used=48.00KiB
Metadata, single: total=2.00GiB, used=1.41GiB
GlobalReserve, single: total=339.00MiB, used=0.00B

P.S. only in openSuse I have found this restriction. How can I disable it? Usually I wrote full path/sudo

Absolute path to 'btrfs' is '/usr/sbin/btrfs', so running it may require superuser privileges (eg. root)

UPD: I made a config and it’s empty here


sudo snapper create-config /
snapper -c root list
single | 0 |                           |      | root         |         | current  |  

some catalogs are not permitted with the command


sudo btrfs fi | du -h --max-depth=1 /

Ok, after 2 weeks I’m planning to upgrade my PC since 6-years of it’s stable work and reinstall the OS (sad, Leap 15.1 still is not ready to the date). So, some gossips made me an idea I shouldn’t use BTRFS next time. But what about XFS? Or still old plain ext4 is better? I’m using really fast SSD and the plan is to upgrade it even faster.

P.S. Oops, my current version is 15.0, need to change topic :slight_smile:

That’s entirely your choice. It really depends on how you use the system.

But what about XFS? Or still old plain ext4 is better?

I have been using “ext4”, and it has not caused problems. I have KaOS running with “xfs”, and that seems fine. But I don’t use it very much, so probably not a good test.

I do have 15.1 running with “btrfs” for testing. But I will be sticking to “ext4” for my main machines.

I also have recommendations:

  • ext4
    for just about everything — ext4 is incredible robust and well-tested (think about many, many Linux-server and -applicance and Android installations alone) and has predictable behavior; for example, commands like »df« provide exact results, as opposed to only roughly approximate results with btrfs, last I checked.
  • qdirstat
    by Stefan Hundhammer who works at SuSE in Nuremberg. He initially developed it as »kdirstat« and renamed it with the advent of KDE/Plasma5, I think. His program provides excellent, intuitive diagrams about what kind of files claim how much disk space, where and in which subdirectories. It can display the whole contents of disks when run as root (without root privileges, you can only see as far as your permissions allow you to). I’ve been using it since I’ve known Stefan, which may be almost two decades now.

Yeah, nice program, I used something like that before, don’t remember. It showed me total used space is 105Gb.

My system was failed to boot in GUI mode because of lack of free space. So, I made an upgrade from a USB, some packets were recycled and now I have 130 GB of free space