Infrastructure & Operations › Linux & Servers

Archiving and Compression (tar, gzip, zip)

Bundling and compressing files.

Also known as: tar, gzip, zip, tar.gz, tarball

Two separate jobs that are often combined:

• Archiving bundles many files into one file. tar does this.
• Compression makes data smaller. gzip does this (to a single stream of data).

A .tar.gz (or .tgz) is an archive that has then been compressed. zip does both in one format and is common when sharing with Windows users.

tar -czf backup.tar.gz project/      # create (c), gzip (z), to file (f)
tar -tzf backup.tar.gz               # list contents (t) without extracting
tar -xzf backup.tar.gz               # extract (x)
tar -xzf backup.tar.gz -C /tmp/out   # extract into another directory

gzip big.log                         # replaces big.log with big.log.gz
gunzip big.log.gz                    # reverse

zip -r site.zip site/                # zip a directory
unzip site.zip

Memory aid: create, xtract, tist, z = gzip, f = the filename comes next. Put f last in the option group, since it takes the next argument.

Mistakes to avoid

• Extracting into the wrong place. Run tar -tzf first to see whether the archive has a top-level folder or spills many files into the current directory.
• Forgetting -r with zip. Without it, a directory is added empty.
• Only compressing what compresses. Text, logs and JSON shrink a lot; JPEGs, videos and already-compressed archives barely shrink.
• Trusting archives from strangers. Look at the paths inside before extracting.

Compression trades CPU time for size, so it is useful for backups, log rotation and transfers over slow links. See scp and rsync for moving the result.