File and Directory Compression

Compression in Linux typically involves packing collections of files into an archive using the tar command, which gets its name from tape archive. Files or directories can be packed into a single tar file, as well as compressed further either the -z option to tar or other programs. The -c flag is used to create an archive and the -x flag is for extracting an archive. The -v option enables verbose output, and -f specifies to store as an archive file. By default, directories are added recursively, unless otherwise specified. Here is an example of creating an archive or tar file:

$ tar -cvf my_archive.tar file1 file2 file3
file1
file2
file3

And to extract the same archive (not verbose):

$ tar -xf my_archive.tar

A program commonly used along with tar is gzip, which creates archives with the extension .gz. A file can be compressed simply by gzip file (with an added -r for a directory) or a .tar.gz file can be created (or extracted) by adding the -z option to a tar command. For example, the same command from above to extract with gzip:

$ tar -xzf my_archive.tar.gz

Another common extension for a gzipped tar file is .tgz.

For more on compression, see this detailed article.

Symbolic Links

Symbolic links are a special type of file which refer to another file in the filesystem. The symbolic link contains the location of the target file. Symbolic links are used to provide pointers to files in more than one place and can be used to facilitate program execution, make navigating on the system easier, and are frequently used to manage system library versions. To make a symbolic link:

$ ln -s data/file/thats/far/away righthere

See the man pages for ln for more information on linking files.

Root and Sudo

The root user on any system is the administrative account with the highest level of permissions and access. This account is sometimes referred to as the superuser. By default, most Linux systems have a single root account when installed and user accounts have to be set up. The root account has a UID of 0, and the system will treat any user with a UID of 0 as root.

If you have access to a root account on any Linux system, best practice is to only use this account when the privileges are needed to perform your work (such as installing packages), and to use a user account for all of your other work. Note that the root directory is not the home directory of the root user, but rather the root of the filesystem. The home directory of the root user is actually located at /root.

The program sudo allows users to run commands with the equivalent privileges of another user. The default privileges selected are the root user's, but any user can be selected. A user with sudo privileges can run commands with root privileges without logging in as root (must enter user's password) by putting sudo in front the command. The first user account created on some Linux distributions is given sudo privileges by default, but most distributions require you to specifically give sudo privileges to a user. This is typically done by editing the /etc/sudoers file (requires either root or sudo access), or running a command like usermod.

Package Managers

The root user and any user with sudo privileges have full access to the features of a package manager. In short, packages are archives of software and associated data, and a package manager is used to install, uninstall, and manage packages on a system. They are used in the shell or through a GUI, and have varying features. Most Linux distributions have a default package manager installed with the system. Some common package managers available are:

Commands for these package managers can be found in their supporting documentation or via the man pages. Note that on a managed resource, the availability of user software is often managed through the Module Utility.

Mounting Storage Volumes or Devices

The mount command can be used to attach the filesystem of another device at a specified place in the directory tree for easy read/write access. mount with no arguments is useful for seeing what devices are mounted. Typically, you must specify the type of the filesystem, name of the device, and the path to where you want to mount it:

mount -t [type] [device] [path]

Use the umount command to unmount a device's filesystem. It has similar options to mount, and both commands have thorough man pages. Another way to mount a device is to use fstab, which automates the process. Network shares can be mounted as well, so long as appropriate credentials are supplied to connect.

 
©  |   Cornell University    |   Center for Advanced Computing    |   Copyright Statement    |   Inclusivity Statement