Linux Basics 2 · Lesson 5 · 30 min

Archives & compression

Downloads, backups and log files arrive as .zip, .tar.gz, .tar.xz, .gz… Two different jobs hide behind those names: archiving (many files → one file) and compressing (one file → a smaller file). Once you can tell them apart, every download makes sense, and you'll never again unpack 200 files straight into your home folder.

You will learn

Archive vs compress, and reading names like .tar.xz
gzip, xz and bzip2: speed vs size, -k, zcat/xzcat
tar: create (c), list (t), extract (x), with z/J/j and -C
zip/unzip for sharing with Windows and Mac
file: what is this really? And how to spot a "tarbomb" before it goes off

Two jobs, many names

Name	What it is	Unpack with
`.tar`	archive only: many files glued together, not smaller	`tar -xf`
`.gz`	ONE file compressed with gzip	`gunzip` (or `gzip -d`)
`.xz` / `.bz2`	ONE file compressed with xz / bzip2	`unxz` / `bunzip2`
`.tar.gz` `.tgz`	a tar archive, then gzipped	`tar -xf`
`.tar.xz` `.tar.bz2`	a tar archive, then xz / bzip2	`tar -xf`
`.zip`	archive AND compression in one (Windows' favourite)	`unzip`

Unix tools each do one job, so tar makes the archive and a compressor squeezes it. Modern tar -xf works out the compression by itself when extracting.

Compressing single files

gzip -k big.log        # big.log.gz, and -k keeps the original
xz -k big.log          # big.log.xz: smaller, but slower to make
ls -l big.log*         # compare the sizes
zcat big.log.gz | head   # read it without unpacking (xzcat, bzcat for the others)
gunzip big.log.gz      # back to big.log

Tool	Speed	Size	Where you meet it
gzip	fast	good	everywhere: web servers, log rotation, `.tar.gz`
bzip2	slow	better	older downloads
xz	slowest to pack, quick to unpack	best	kernel sources, packages, software releases

Text (logs, CSV, code) often shrinks to a tenth of its size. Photos, videos and anything already compressed barely shrink at all. Zipping a JPEG is a waste of time.

tar: the archiver

tar -czf backup.tar.gz Documents     # c = create, z = gzip, f = file name (always last before the name!)
tar -cJf backup.tar.xz Documents     # J = xz   (j = bzip2)
tar -tf backup.tar.gz                # t = list what's inside, without unpacking
tar -xf backup.tar.gz                # x = extract, here
tar -xf backup.tar.gz -C /tmp/restore   # …into another folder
tar -caf site.tar.xz site/           # a = pick the compression from the name

Add v (verbose) to see file names fly by: tar -xvf. The Backups lesson uses tar for real backups: Linux Sysadmin, lesson 8.

Tarbombs

A well-made archive has one folder at the top. A tarbomb has dozens of loose files that explode into whatever folder you're in. Always list first: tar -tf x.tar.gz | head. If there's no single top folder, make one and extract into it: mkdir x && tar -xf x.tar.gz -C x. Same for zip: unzip -l first, then unzip file.zip -d folder.

zip for everyone else

zip -r homework.zip Documents    # -r = include folders' contents
unzip -l website.zip             # list
unzip website.zip -d site        # extract into a folder

Rocky / RHEL

Minimal installs have gzip, xz and tar, but not zip, unzip or bzip2: sudo dnf install zip unzip bzip2.

Ubuntu / Debian

gzip, xz, bzip2 and tar are there. sudo apt install zip unzip for zip files.

What is this file, really?

file downloads/*

downloads/mystery:      gzip compressed data, was "treasure.txt", from Unix
downloads/website.zip:  Zip archive data, at least v2.0 to extract
downloads/logs.tar.xz:  XZ compressed data, checksum CRC64

Linux doesn't care about extensions. file reads the first bytes (the "magic number") and tells you what's really inside, even when the name lies or is missing.

Practice: unpack the downloads folder 📦

~/downloads has a zip, some tarballs, a gzipped note, a file with no extension at all, and one archive that's a trap. There's also a big log in your home folder to squeeze.

Quick check

1. What's inside photos.tar.gz?

2. Before extracting an archive from the internet, what do you run?

3. You gzip a folder of JPEG photos. Why is the result almost the same size?

Finished the missions and the quiz? Mark it done to track your progress.