My disk space is dwindling by about 2GB a day! I only have a few more days before I run out of space.

$ df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/sda4             143G  126G   11G  93% /
udev                  491M  4.0K  491M   1% /dev
tmpfs                 200M  696K  199M   1% /run
none                  5.0M     0  5.0M   0% /run/lock
none                  499M  144K  499M   1% /run/shm
/dev/sda2             1.9G  580M  1.2G  33% /tmp
/dev/sda1              92M   29M   58M  33% /boot

I have been searching for the biggest directories/log files, deleting and compressing. But I am still losing the war. Finally, I realised I have a big misunderstanding:

julian@server1:~$ sudo du -h / | tail -n 1
16G     /

All of my files in / only add up to 16 GB. That leaves 110 GB unaccounted for!

Clearly I have a misunderstanding: I thought the '/dev/sda4' line represented all the files visible from '/'. What should I be reading to understand where the other storage has gone?

More details:

• I have an Ubuntu 11.10 server, that was set-up by data-center staff.

• It is running

  • my own code (which is fairly prolific with log files, but otherwise doesn't store much stuff on the drive)
  • duplicity for backups (which tends to store a lot of signature files)
  • various other standard services, like Apache, nagios, etc. They are very lightly used.

• It has been up for about 4 months without a reboot.

• I lied about the du output (simplified it for effect). It also complained about not being able to access GVFS and the du processes's own resources. I believe they are irrelevant:

.

 du: cannot access `/home/julian/.gvfs': Permission denied
 du: cannot access `/proc/10841/task/10841/fd/4': No such file or directory
 du: cannot access `/proc/10841/task/10841/fdinfo/4': No such file or directory
 du: cannot access `/proc/10841/fd/4': No such file or directory
 du: cannot access `/proc/10841/fdinfo/4': No such file or directory