I have the following Debian that works in Virtual Box:

$ uname -a
Linux debian 3.2.0-4-amd64 #1 SMP Debian 3.2.51-1 x86_64 GNU/Linux

Recently I notice that I don't have any free space:

$ df
Filesystem                                             1K-blocks    Used Available Use% Mounted on
rootfs                                                   1922060 1921964         0 100% /
udev                                                       10240       0     10240   0% /dev
tmpfs                                                     206128     296    205832   1% /run
/dev/disk/by-uuid/ef55765f-dae5-426f-82c4-0d98265c5f21   1922060 1921964         0 100% /
tmpfs                                                       5120       0      5120   0% /run/lock
tmpfs                                                     511980       0    511980   0% /run/shm
/dev/sda3                                                5841936  163548   5381636   3% /home
tmpfs                                                     511980      12    511968   1% /tmp

What is the /dev/disk/by-uuid/ef55765f-dae5-426f-82c4-0d98265c5f21? Why does it use all free space on the disk?

  • 1
    That's probably the device that contains your root filesystem, because the mountpoint listed there is /. You can check what device that UUID points to by running ls -l /dev/disk/by-uuid/ef55765f-dae5-426f-82c4-0d98265c5f21. Oh, and apparently your root filesystem is full, so you should check what's using all your space ;) – Martin von Wittich Jan 29 '14 at 9:15
  • 1
    @Martin von Wittich How can I check it? I'm new to the Linux world – FrozenHeart Jan 29 '14 at 9:19
  • 2
    du -sh calculates the size of a given folder. Run du -sh /* as root to determine which folder is the biggest, then cd into that folder, and run du -sh *. Repeat until you've found the folder that's using up all your space. – Martin von Wittich Jan 29 '14 at 10:08
  • 1
    There's also durep which makes a nice recursive graph (durep -td 4 -hs 100M /), but that's usually not installed by default and so currently of little use to you, because you probably won't be able to install new packages until this problem is resolved. – Martin von Wittich Jan 29 '14 at 10:09

That device has the same blocks, used and free space as your rootfs filesystem, so they are probably the same. You can check where the uuid points to with:

ls -l /dev/disk/by-uuid/ef55765f-dae5-426f-82c4-0d98265c5f2

My guess is that you just booted from a life filesystem on a CDROM image.


The device /dev/disk/by-uuid/ef55765f-dae5-426f-82c4-0d98265c5f21 is simply a symbolic link to an actual device.

The reason for using links is because kernel device names (e.g. /dev/sda, /dev/sdb) may change depending on where or when the disk is plugged in, whereas the links in /dev/disk/by-* will always point to the same drive no matter what, and therefore are safer to use.


In modern Linux systems whatever you see under /dev (and a few other directories like /proc and /sys) is fake, it is a view into kernel internals or the machine configuration). In original Unix systems under /dev by convention you'd find the special files corresponding to the devices configured into the system.


Since your seem to run Debian Wheezy, you can look at /etc/fstab. Above the line that describes the root partition is a line like this:

# / was on /dev/sda1 during installation

Modern motherboards tend to enumerate SATA disks differently if you add or remove one, so the debian installer writes a comment about the device descriptor in use at installation time, which may not be the same anymore.

As Tim said, you can get the actual device with

ls -l /dev/disk/by-uuid/ef55765f-dae5-426f-82c4-0d98265c5f2

Assuming this is a real hard disk installation, it can also be full because deleted files are not yet seen as free space: for example, if you just upgraded some packages used by programs actually running, then the file --eg libc6.so-- is replaced by the new one, but the inode of the old one is still in use. apt-get clean && reboot is your friend in this case.


I have the same recurring problem, which I solve by doing :

rm -rfv /var/log/{*.[0-9],*.[0-9].gz}
for i in `find /var/log/ -type f`; do
        if [ -f $i ] && [ -s $i ]; then
                echo "" > $i

This I got from : https://serverfault.com/questions/350891/how-to-delete-all-debian-squeeze-logs?rq=1 And rebooting BUT I don't understand WHY !

I've stopped this recurring by deleting these files found "suspected" by my anti-virus ClamTk :

/home/daniel/.cache/chromium/Default/Cache/f_03b482 PUA.JS.Xored
/home/daniel/.cache/chromium/Default/Cache/f_03b53f PUA.HTML.Infected.WebPage-1
/home/daniel/.cache/chromium/Default/Cache/f_03b547 PUA.Phishing.Bank
/home/daniel/.cache/chromium/Default/Cache/f_03b61d PUA.JS.Xored
/home/daniel/.cache/chromium/Default/Cache/f_03b65a PUA.Script.Packed-2

So these were files in the cache of one of my web browsers. And I remember that this problem started in a period when my firewall "Firestarter" stopped a bunch of incoming connections and I was using Chromium.

Did my case help anyone ?

  • @nwildner thanks for the edit, but please don't add "Edit" to questions or answers here. Posts on SE sites should be timeless, there's no need to explicitly mention whether they've been edited. They simply need to make sense when someone reads them for the first time. – terdon Jan 14 '16 at 12:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.