Today the hard drive of our server was suddenly full. The disk usage always stayed around 50 % in the weeks and months before (old data is regularly expunged from the server).

I deleted 10 GB of files in /tmp, which strangely freed 51 GB. Here is what I did:

root@***:~# df -h
Dateisystem           Size  Used Avail Use% Eingehängt auf
/dev/sda3             139G  137G     0 100% /
tmpfs                 3,9G     0  3,9G   0% /lib/init/rw
udev                  3,9G  116K  3,9G   1% /dev
tmpfs                 3,9G     0  3,9G   0% /dev/shm
/dev/sda1             985M   25M  910M   3% /boot

root@***:/var# du -hs *
3,3M    backups
438M    cache
9,4G    lib
4,0K    local
12K     lock
76M     log
24K     mail
4,0K    opt
88K     run
184K    spool
10G     tmp
12K     www

root@***:/var/tmp# find -type f -print0 | xargs -0 rm


root@***:/var/tmp# df -h
Dateisystem           Size  Used Avail Use% Eingehängt auf
/dev/sda3             139G   81G   51G  62% /
tmpfs                 3,9G     0  3,9G   0% /lib/init/rw
udev                  3,9G  116K  3,9G   1% /dev
tmpfs                 3,9G     0  3,9G   0% /dev/shm
/dev/sda1             985M   25M  910M   3% /boot

Any explanation as to why deleting 10 GB in /tmp gave me back 51 GB on the disk? Could this point to an SSD failure? Are there any tools for Debian to test SSD health?

I already have checked syslog. The first entry relating to this incidient is a mysql message:

1:22:02 [ERROR] /usr/sbin/mysqld: Disk is full writing...

So I have absolutely no idea what caused this.