How to clean up unnecessary files

Question

Please suggest me any particular unnecessary file that I can clean to back everything to normal condition(temporarily). (i.e. any log or archieve or anything ). My var/log has only 40MB and Home directory has 3GB of space(so I believe that's not a problem). Other than that what I can clean up to make space.

[user@host]$ df -h
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/vg_inamivm-lv_root
 18G 17G 0 100% /
tmpfs 1.9G 0 1.9G 0% /dev/shm
/dev/sda1 485M 71M 389M 16% /boot

I am in a debian machine.

UPDATE1:

output of cd /; du -sxh *

6.1M bin
61M boot
156K dev
22M etc
3.3G home
306M lib
18M lib64
16K lost+found
4.0K media
4.0K mnt
408K opt
du: cannot access `proc/18605/task/18605/fd/4': No such file or directory
du: cannot access `proc/18605/task/18605/fdinfo/4': No such file or directory
du: cannot access `proc/18605/fd/4': No such file or directory
du: cannot access `proc/18605/fdinfo/4': No such file or directory
0 proc
208K root
9.7M sbin
0 selinux
4.0K srv
0 sys
8.0K tmp
536M usr
187M var

Update2

Output of ls -la /

dr-xr-xr-x. 22 root root 4096 Aug 7 08:42 .
dr-xr-xr-x. 22 root root 4096 Aug 7 08:42 ..
-rw-r--r--. 1 root root 0 Aug 7 08:42 .autofsck
dr-xr-xr-x. 2 root root 4096 Mar 28 16:53 bin
dr-xr-xr-x. 5 root root 1024 Mar 28 16:54 boot
drwxr-xr-x. 16 root root 3580 Sep 9 03:13 dev
drwxr-xr-x. 69 root root 4096 Aug 23 09:19 etc
drwxr-xr-x. 9 root root 4096 Jun 29 16:10 home
dr-xr-xr-x. 8 root root 4096 Mar 7 2012 lib
dr-xr-xr-x. 9 root root 12288 Mar 28 16:53 lib64
drwx------. 2 root root 16384 Mar 7 2012 lost+found
drwxr-xr-x. 2 root root 4096 Sep 23 2011 media
drwxr-xr-x. 2 root root 4096 Sep 23 2011 mnt
drwxr-xr-x. 3 root root 4096 Mar 7 2012 opt
dr-xr-xr-x. 355 root root 0 Aug 7 08:42 proc
dr-xr-x---. 5 root root 4096 Aug 17 18:27 root
dr-xr-xr-x. 2 root root 4096 May 2 09:13 sbin
drwxr-xr-x. 7 root root 0 Aug 7 08:42 selinux
drwxr-xr-x. 2 root root 4096 Sep 23 2011 srv
drwxr-xr-x. 13 root root 0 Aug 7 08:42 sys
drwxrwxrwt. 3 root root 4096 Sep 13 03:37 tmp
drwxr-xr-x. 13 root root 4096 Mar 28 17:53 usr
drwxr-xr-x. 18 root root 4096 Mar 7 2012 var

• Krzysztof Adamski: There is no universal answer here. Run following commands as root, wait for a while and paste your results: cd /; du -sxh *. It will sum up the size of all main directories on your server. Then, we will have a clue. You can also enter this big directory with cd and run du command inside of it to get deeper.
• Subhransu Mishra: @KrzysztofAdamski Thanks for a quick reply. I have updated the question with output of cd /; du -sxh *
• Karlson: @Subhransu Based on what you are seeing it appears as if you have a file that has been removed while still open.
• Subhransu Mishra: @Karlson I really don't understand if you have a file that has been removed while still open . Please explain. Please let me know what other output could be viable so that you can help me.
• Krzysztof Adamski: Did you remove any big files recently? Like log files or something like this?

Answer 1

daisy's answer to use a graphical tool to visually find large files and directories is probably the best method. However, do note that "graphical tool" does not mean "requires an X server"! The wonderful ncdu program provides the graphical output in the CLI, and works perfectly on remote servers via SSH:

$ ncdu /
. 43.7GiB [##########] /home
. 5.9GiB [# ] /usr
 1.1GiB [ ] /lib
. 1.1GiB [ ] /var
 736.9MiB [ ] /opt
. 324.6MiB [ ] /tmp
 218.4MiB [ ] /boot
. 63.8MiB [ ] /etc
 10.0MiB [ ] /sbin
 8.8MiB [ ] /bin
 3.3MiB [ ] /lib32
. 1.0MiB [ ] /run
 64.0KiB [ ] /build
! 16.0KiB [ ] /lost+found
 8.0KiB [ ] /media
 8.0KiB [ ] /mnt
 8.0KiB [ ] /.config
 4.0KiB [ ] /dev
 4.0KiB [ ] /lib64
e 4.0KiB [ ] /srv
e 4.0KiB [ ] /selinux
! 4.0KiB [ ] /root
e 4.0KiB [ ] /cdrom
. 0.0 B [ ] /proc
. 0.0 B [ ] /sys
@ 0.0 B [ ] initrd.img.old
@ 0.0 B [ ] initrd.img
@ 0.0 B [ ] vmlinuz.old

Then, after entering /var/ for instance:

. 395.3MiB [##########] /tmp
. 365.0MiB [######### ] /cache
. 297.8MiB [####### ] /lib
 16.1MiB [ ] /backups
. 8.0MiB [ ] /log
. 56.0KiB [ ] /spool
 40.0KiB [ ] /games
 8.0KiB [ ] /www
e 4.0KiB [ ] /opt
e 4.0KiB [ ] /mail
e 4.0KiB [ ] /local
e 4.0KiB [ ] /crash
@ 0.0 B [ ] lock
@ 0.0 B [ ] run

Install easily on Debian or Ubuntu:

$ sudo apt-get install ncdu

Install easily on CentOS as root:

# yum install ncdu

• annahri: Oh wow, didn't know ncdu existed. All this time, I use du -sh ./* 2> /dev/null | sort -rh.

Answer 2

The best way of finding out disk consuming, is using graphical software like baobab:

Launch it with sudo baobab /

enter image description here

• Subhransu Mishra: Will it work in remote server ?
• Gilles 'SO- stop being evil': @Subhransu Sure, if you connect with X forwarding (ssh -X or ForwardX11 yes in ~/.ssh/config). However, if it isn't installed you might not have enough space to install it and the requisite directory.
• dotancohen: See my answer below for a graphical tool that does not require an X server and requires very little server space.
• mYnDstrEAm: gdmap (Graphical Disk Map) can also be used on Debian. It shows a treemap for the entire directory/drive at once. Also GUIs shouldn't be run with sudo.

Answer 3

Debian offers a utility called cruft IIRC that lists files that could possibly be uneeded on your system.

• nilon: How can that be used?
• mYnDstrEAm: After installing cruft-ng when running it or cruftit says warning: plocate database is outdated. Using KDE which uses baloo and don't want to create some large database for long just for that.

Answer 4

make a new file in /bin called treesize

chmod +x /bin/treesize

paste this in it.

#/bin/sh
du -k --max-depth=1 | sort -nr | awk '
 BEGIN {
 split("KB,MB,GB,TB", Units, ",");
 }
 {
 u = 1;
 while (1ドル >= 1024) {
 1ドル = 1ドル / 1024;
 u += 1
 }
 1ドル = sprintf("%.1f %s", 1,ドル Units[u]);
 print 0ドル;
 }
 '

Output looks like this.

#treesize
3.0 GB .
1.1 GB ./usr
759.9 MB ./var
353.3 MB ./root
307.1 MB ./opt
270.7 MB ./lib
98.6 MB ./home
60.1 MB ./boot
9.8 MB ./etc
8.3 MB ./bin
8.1 MB ./sbin
268.0 KB ./run
40.0 KB ./tmp
16.0 KB ./lost+found

• Alejandro: or this du -hsx /* | sort -rh | head -n 40
• Seamus: @Alejandro: it's not really the same - at all.
• Seamus: I like this - very neat!
• Alejandro: it works on my machine

Answer 5

You can check for deleted files with lsof | grep -i deleted Then you can see if a process is hanging on to an inode that you think was deleted. If so, restart the parent process to release the old (deleted) file.

Answer 6

I've found some very useful commands on this post at Askubuntu.com. Paraphrasing:

• Show top 10 biggest subdirs in the current dir: du -sk * | sort -nr | head -10
• Use filelight/kDirStat/baobab to see where the disk space is going visually
• Check if you have old kernels for deletion: ls -lh /boot
• Clean packages: sudo apt-get autoremove and sudo apt-get autoclean
• See list of all installed packages, sorted by size: dpkg-query -W --showformat='${Installed-Size} ${Package}\n' | sort -nr | less
• Clean unused language files with translations: sudo apt-get install localepurge
• Check content of /var/tmp/: du -sh /var/tmp/
• Check also man deborphan
• Search for big files: find / -type f -size +1024k or find / -size +50000 -exec ls -lahg {} \;

Answer 7

This will give you the biggest files on your FS and maybe you will find there some files which can be deleted.

find / -xdev -type f -size +100000c -exec ls -la {} \; 2>/dev/null | sort -nk5 | tail -20

If you need longer output, just change number after tail command.
Please post output from:

df -h /
du -shx /

Both commands should show similar used space of your FS.

Answer 8

find . -atime +180 | xargs ls -al | sort -u -n -k5 | numfmt --to=iec --field=5

• find all files in . whos access time is over 180 days old
• show file size
• sort by 5th column, numerical
• format 5th column to human readable

Useful links about access time:

• https://superuser.com/questions/464290/why-is-cat-not-changing-the-access-time
• timestamp, modification time, and created time of a file