Df and du are two very useful utilities which are normally installed by default in all Linux distributions. We can use the first one to obtain an overview of the used and available space on mounted filesystems; the second, instead, is very useful to obtain a detailed report about the space used by files and directories. In this article we take a look at their usage, and we see what are the most common used options which can be used to modify their behavior.
In this tutorial you will learn:
- How the df utility works
- How to show the output of df in human-readable form
- How to include the filesystem type in the output of df
- How to include or exclude filesystems from the output of df
- How the du utility works
- How to obtain a human-friendly output with du
- How to obtain a summary of the used space
- How to exclude files from the output of du
- How to obtain a “grand total” of the space in use by multiple directories
Software requirements and conventions used
|Category||Requirements, Conventions or Software Version Used|
|Software||df and du (installed by default)|
|Other||No other requirements needed|
|Conventions||# – requires given linux-commands to be executed with root privileges either directly as a root user or by use of
$ – requires given linux-commands to be executed as a regular non-privileged user
df utility is used to display, among the other things, filesystems available and used disk space. If the program is called without any argument, all mounted filesystem are included in the report:
$ df Filesystem 1K-blocks Used Available Use% Mounted on devtmpfs 2908316 0 2908316 0% /dev tmpfs 2930156 28392 2901764 1% /dev/shm tmpfs 2930156 1792 2928364 1% /run tmpfs 2930156 0 2930156 0% /sys/fs/cgroup /dev/mapper/fingolfin_vg-root_lv 35862048 7210616 26800040 22% / tmpfs 2930156 124 2930032 1% /tmp /dev/sda1 1032088 161664 817996 17% /boot /dev/mapper/fingolfin_vg-home_lv 25671908 1515396 22829408 7% /home /dev/mapper/fingolfin_vg-data_lv 152737296 90208196 54700828 63% /mnt/data /dev/dm-5 152786272 90208644 54746804 63% /mnt/databk tmpfs 586028 1124 584904 1% /run/user/1000
Let’s take a look at the output of the command. In the first column we have the filesystem path, in the second we have its size, which is reported in 1K blocks. In the third and fourth columns we have the used and available space respectively. In the fifth column the usage percentage is reported, while in the last column we can see the filesystem mountpoint.
Df accepts one or more file paths as arguments. When those arguments are provided only the information about the filesystem on which the files reside will be reported. For example, if we specify
/home/egdoc/.bashrc as file argument, we will get a report about the filesystem mounted on
/home, which, in this case, is on a separate LVM logical volume:
$ df /home/egdoc/.bashrc Filesystem 1K-blocks Used Available Use% Mounted on /dev/mapper/fingolfin_vg-home_lv 25671908 1515468 22829336 7% /home
If we pass a mountpoint as argument to the command, information about the mounted filesystem are reported instead.
Displaying information in human-readable form
As you can see in the output of the previous examples, filesystem sizes and used space are reported in 1K-blocks, so they are not very human friendly. To get those information in a format more understandable by us humans, we can invoke
df with the
-h option, which is the short for
--human-readable. When the option is provided, the sizes are reported in power of
1024, with a convenient suffix:
$ df -h /home/egdoc/.bashrc Filesystem Size Used Avail Use% Mounted on /dev/mapper/fingolfin_vg-home_lv 25G 1.5G 22G 7% /home
Including or excluding filesystems
In some cases we may want only some specific filesystems to be included in the output of the command. We will see how to achieve this task in a moment, but first let’s see how to include the filesystem type in the output of
--print-type) option let us achieve exactly that:
$ df -h -T Filesystem Type Size Used Avail Use% Mounted on devtmpfs devtmpfs 2.8G 0 2.8G 0% /dev tmpfs tmpfs 2.8G 22M 2.8G 1% /dev/shm tmpfs tmpfs 2.8G 1.8M 2.8G 1% /run tmpfs tmpfs 2.8G 0 2.8G 0% /sys/fs/cgroup /dev/mapper/fingolfin_vg-root_lv ext4 35G 6.9G 26G 22% / tmpfs tmpfs 2.8G 124K 2.8G 1% /tmp /dev/sda1 ext2 1008M 158M 799M 17% /boot /dev/mapper/fingolfin_vg-home_lv ext4 25G 1.5G 22G 7% /home /dev/mapper/fingolfin_vg-data_lv ext4 146G 87G 53G 63% /mnt/data /dev/dm-5 ext4 146G 87G 53G 63% /mnt/databk tmpfs tmpfs 573M 1.2M 572M 1% /run/user/1000
To exclude or include certain filesystem types from the output of the command, we can basically use two options:
--exclude-type). The option names are pretty self explanatory: we can use the former to specify the
filesystem type which should be included in the output. For example, to get information only about
ext4 filesystems, we would run:
$ df -h -t ext4 Filesystem Size Used Avail Use% Mounted on /dev/mapper/fingolfin_vg-root_lv 35G 6.9G 26G 22% / /dev/mapper/fingolfin_vg-home_lv 25G 1.5G 22G 7% /home /dev/mapper/fingolfin_vg-data_lv 146G 87G 53G 63% /mnt/data /dev/dm-5 146G 87G 53G 63% /mnt/databk
The latter option,
-x, let us specify the filesystem type which should be excluded from the output of
df, instead. To display all filesystems except the
ext4 ones, we would run:
$ df -h --x ext4 Filesystem Size Used Avail Use% Mounted on devtmpfs 2.8G 0 2.8G 0% /dev tmpfs 2.8G 22M 2.8G 1% /dev/shm tmpfs 2.8G 1.8M 2.8G 1% /run tmpfs 2.8G 0 2.8G 0% /sys/fs/cgroup tmpfs 2.8G 124K 2.8G 1% /tmp /dev/sda1 1008M 158M 799M 17% /boot tmpfs 573M 1.2M 572M 1% /run/user/1000q
du utility is used to report directories and file space usage. When invoked with no arguments, the utility (recursively) displays the list of all the directories existing in the current position, together with their space usage (the output below is truncated for convenience):
$ du 4 ./Public [...] 4 ./Documents 4 ./Videos 4 ./Pictures 1469376 .
Just like in the case of
df, we can make the
du command produce a more human-readable output by using the
-h option, short for
--human-readable. When the option is used, for example, the output above becomes:
$ du -h 4.0K ./Public [...] 4.0K ./Documents 4.0K ./Videos 4.0K ./Pictures 1.5G .
In case we want to visualize only the summary of the used disk space in a certain directory, we can use the
-s option, which is the short for
--summarize. If we invoke the command with this option, in the same working directory as before, we obtain the following result:
$ du -hs 1.5G .
Including files and their size in the output of du
By default, as we can observe, only directories are reported in the output of the
du command. To include also standard files in the output, we can use the
-a option (short for
-all). If we re-launch the command with this option we obtain the following:
$ du -ha 4.0K ./Public 4.0K ./.Xauthority 4.0K ./.bashrc [...] 4.0K ./Documents 4.0K ./Videos 4.0K ./Pictures 1.5G .
As you can see, the
.bashrc files (and the respective space usage), were included in the (truncated) output.
Excluding files matching a pattern
We have basically two ways to specify the files and directories which should be excluded from the output of the
du command: the first one is by invoking the program with the
--exclude option. This option takes a pattern as argument; all the files and directory names matching that pattern will be excluded. Just as an example, suppose we want to exclude all hidden files from the output of
du. We know hidden files are named starting with a dot, therefore we could run:
$ du -ha --exclude=./.* 4.0K ./Public 4.0K ./Music 4.0K ./Downloads 4.0K ./Desktop 4.0K ./Templates 4.0K ./Documents 4.0K ./Videos 4.0K ./Pictures 36K .
As expected, only non-hidden files and directories were included in the output. We can specify multiple exclusion patterns by repeating the
--exclude option. For example, suppose we want to also exclude files and directories which
have a name starting with the
D character. We would run:
$ du -ha --exclude=./.* --exclude=./D* 4.0K ./Public 4.0K ./Music 4.0K ./Templates 4.0K ./Videos 4.0K ./Pictures 24K .
The second way we can use to specify pattern that should be used for file and directory exclusions, is to use the
-X option (
--exclude-from). This option takes the name of a file as argument: is in that file that we specify the patterns to be matched. Let’s create it. We will call it
$ $ cat << EOF > exclusions.txt > ./.* > ./D* > EOF
With our file in place we can run:
$ du -ha --exclude-from=exclusions.txt 4.0K ./Public 4.0K ./Music 4.0K ./exclusions.txt 4.0K ./Templates 4.0K ./Videos 4.0K ./Pictures 28K .
As you can see we obtained the same result as above, except for the fact that the size of the exclusions file (
exclusions.txt) is now included in the count of the space used by the current directory (the last line of the output:
Obtaining a “grand total” of the used space
du utility accepts multiple files and directories as its arguments. For example we can use it to obtain a summary of the space used by the
$ sudo du -hs /etc /boot 28M /etc 157M /boot
In the example above we ran du with
sudo in order to grant it access to some restricted directories. As you can see the summary of the space usage is reported for each directory we specified. What if we want to obtain a “grand total”, so to have the sum of the space used by the two directories? To achieve that we can use the
-c option (short for
--total). Our command becomes:
$ sudo du -hsc /etc /boot 28M /etc 157M /boot 184M total
In this article we learned to use
du, two utilities installed by default in all Linux distributions. They are very useful to check the available and used space in mounted filesystems, and to obtain detailed information about the size of file and directories. We covered the most common use cases; for more detailed information, as always, check the manuals!