tar Command in Linux: Pack and Unpack Archives

tar bundles many files into one archive — originally for tape drives, today for backups, source drops, and release artefacts. With compression it is the .tar.gz you meet on every download page.

The two commands everyone memorises #

Create and extract a compressed archive

$ tar -czf project.tar.gz project/
$ tar -xzf project.tar.gz

Read the letters as a sentence: c create, z gzip-compress, f filename; x extract. The order of the flags does not matter; what matters is choosing c or x, and that f comes last before the filename.

The full flag set #

Flag	Meaning
`-c`	Create an archive
`-x`	Extract an archive
`-t`	List contents without extracting
`-z`	Filter through gzip (`.tar.gz`); `-j` bzip2, `-J` xz
`-f`	The archive filename, always last
`-v`	Verbose: list files as they are processed
`-C dir`	Change to dir before acting
`--exclude=P`	Skip paths matching a pattern

Verify before you extract #

List the contents of an archive

$ tar -tzf project.tar.gz | head

One habit prevents the classic mistake — an archive that explodes 40 files into your current directory: list first. If the entries lack a leading folder, extract into a new one:

Extract into a directory you control

$ mkdir restore && tar -xzf project.tar.gz -C restore

Backups with excludes #

Back up a project, skipping caches

$ tar -czf backup.tar.gz --exclude="node_modules" --exclude=".git" project/

Modern GNU tar detects compression from the filename, so tar -xf anything.tar.gz works too — but writing the z explicitly costs nothing and works on older systems.

To move an archive to another machine, pair it with ssh; to publish or fetch one, curl uploads and downloads it. More in the journal.