4.3 Archive and Compress with tar, gzip, and bzip2

4.3 Archive and Compress with tar, gzip, and bzip2

Quick Answer: Create a gzip-compressed tarball with tar czf archive.tar.gz tree/, list with tar tzf archive.tar.gz, extract with tar xzf archive.tar.gz. Use j instead of z for bzip2 (.tar.bz2). Standalone gzip / bzip2 compress single files; tar is for multi-file trees.

The EX200 skill to archive, compress, unpack, and uncompress files using tar, gzip, and bzip2 is the exact wording Red Hat publishes for the RHEL 10 exam. Older revisions of the objective list also named star; the current list does not. This skill appears whenever you backup a tree, ship logs, or unpack vendor content. On RHEL, these tools are always available in a normal base environment—do not depend on exotic compressors unless the task installs them.

Why Archive and Compress Are Separate Ideas

A .tar file is an archive that may still be large. A .gz file is usually a single compressed stream. A .tar.gz (or .tgz) is a tarball compressed with gzip. A .tar.bz2 uses bzip2.

tar Fundamentals

tar originally meant “tape archive.” Today it writes archive files just as often as tapes.

Mode letters you must know

Traditional style concatenates letters after a dash (or even without a dash on some examples):

tar -cf backup.tar project/
tar -tf backup.tar
tar -xf backup.tar

With compression:

tar -czf backup.tar.gz project/
tar -tzf backup.tar.gz
tar -xzf backup.tar.gz

tar -cjf backup.tar.bz2 project/
tar -tjf backup.tar.bz2
tar -xjf backup.tar.bz2

Memory hook: create / xtract / test-list + file + z gzip or j bzip2. The f archive name must immediately associate with the file operand—put the archive filename right after the options that include f (common form: tar czf name.tar.gz ...).

Create examples

cd /home/student
tar czf /root/student-project.tgz project
tar czf app-backup-$(date +%F).tar.gz /srv/app
tar cjf /backup/var-log.tar.bz2 -C / var/log

-C / changes directory before adding members so the archive stores var/log/... rather than an absolute /var/log/... layout—often cleaner for portable restores.

Extract examples

mkdir -p /tmp/restore && cd /tmp/restore
tar xzf /root/student-project.tgz
tar xjf /backup/var-log.tar.bz2

Extract specific members:

tar xzf backup.tar.gz path/inside/archive/file.conf

List before you extract into a sensitive location:

tar tzf backup.tar.gz | head
tar tzf backup.tar.gz | grep 'etc/hosts'

Exam trap: Extracting as the wrong user, in the wrong directory, or overwriting live configs without a listing pass. Always pwd and prefer extracting into an empty staging directory first when unsure.

Absolute paths and tar

Older habits sometimes archived absolute paths. GNU tar often strips leading / on extract by default (safer), but you should still inspect with tar tf so you know whether members look like etc/hosts or /etc/hosts. Do not extract untrusted archives as root over / without understanding contents.

gzip and bzip2 as Standalone Tools

gzip

gzip large.log            # creates large.log.gz and removes large.log by default
gzip -k large.log         # keep original if supported (gzip --keep)
gunzip large.log.gz       # restore
zcat large.log.gz         # view without writing decompressed file
gzip -d large.log.gz      # same idea as gunzip

Characteristics:

• Very common on RHEL for logs and tarballs
• Fast compression and decompression
• Typically smaller gains than bzip2 on some text, but quicker

bzip2

bzip2 bigdump.sql         # creates bigdump.sql.bz2, removes original by default
bzip2 -k bigdump.sql      # keep input when supported
bunzip2 bigdump.sql.bz2
bzcat bigdump.sql.bz2 | less

Characteristics:

• Often better compression ratio than gzip on highly redundant data
• Slower compress/decompress—noticeable on large trees
• Pairs with tar via j → .tar.bz2

Single-file vs tar-of-many

# Wrong mental model for a directory:
gzip mydir                # does not archive a directory tree the way tar does

# Right approach for a directory:
tar czf mydir.tar.gz mydir

gzip/bzip2 operate on files (streams). Directories need tar (or another archiver) first.

Combined Workflows You Will Use on the Exam

Backup a project tree with gzip

tar czf /backup/website-$(date +%F).tar.gz -C /var/www html
ls -lh /backup/website-*.tar.gz
tar tzf /backup/website-*.tar.gz | head

Backup with bzip2 when size matters more than speed

tar cjf /backup/website.tar.bz2 -C /var/www html

Unpack vendor content

cd /usr/local/src
tar xzf /tmp/app-1.2.3.tar.gz
cd app-1.2.3

Compress a single large log without tar

gzip -k /var/log/myapp/old.log
ls -lh /var/log/myapp/old.log*

Pipe tar across a stream (conceptual)

tar czf - project | ssh host 'cat > /backup/project.tgz'
# or extract from stdin on the far side—patterns appear in multi-host labs

Preserving Attributes and Special Files

For ordinary user trees, default tar create/extract is often enough. For system backups you may care about:

• Permissions and ownership (extract as root to restore original owners)
• SELinux contexts (GNU tar options such as --selinux when creating/extracting in SELinux environments)
• Sparse files, device nodes, and ACLs on advanced backups

RHCSA-level tasks usually focus on getting the correct tree back with expected modes. If you archive as root and extract as an unprivileged user into your home directory, ownership becomes yours—verify with ls -l after extract when ownership matters.

cp -a (previous section) is still better for live local clones. Use tar when you need a single portable file, compression, or transfer packaging.

star: legacy objective wording only

Older EX200 objective lists (RHEL 7/8 era) named star (Schily tar) alongside tar. The current RHEL 10 objective names only tar, gzip, and bzip2, and star is not part of a standard RHEL 10 install. Treat any study material that still drills star as out of date for this exam. Master GNU tar + gzip + bzip2; do not spend scarce minutes learning a second archiver that the objective no longer lists.

Verification Checklist After Create/Extract

1. Archive exists and size is plausible: ls -lh backup.tar.gz
2. Contents list looks right: tar tzf backup.tar.gz | head / grep
3. Extract into staging: empty directory, then compare with diff -r or ls -lR
4. Spot-check a known file’s content and mode
5. Clean up temporary copies so you do not leave wrong trees in graded paths

mkdir -p /tmp/tarcheck && cd /tmp/tarcheck
tar xzf /backup/website.tar.gz
find . -type f | head

Common Pitfalls

1. Swapping create and extract — tar xzf on a non-archive path errors; tar czf with wrong argument order can create an archive named after a file you meant to include.
2. Forgetting f — tar may try to use a default tape device.
3. Wrong compression letter — using z on a .bz2 archive (or the reverse) produces errors; list with the matching filter or use modern tar auto-detect forms where available (tar xf file.tar.gz often works on GNU tar).
4. gzip on a directory — does not replace tar.
5. Extracting in / or /etc accidentally — always check pwd and archive member prefixes.
6. Assuming compression keeps originals — classic gzip file removes file unless keep is requested.
7. Broken workflows with symlinks — decide whether you need to store symlinks as links (tar does) or follow them; know that extracting may recreate links that break if targets are outside the archive.

Quick Reference Card

Practice Drill

mkdir -p /tmp/tarlab/src/sub
echo one > /tmp/tarlab/src/a.txt
echo two > /tmp/tarlab/src/sub/b.txt
ln -s a.txt /tmp/tarlab/src/link-to-a
cd /tmp/tarlab
tar czf src-gzip.tar.gz src
tar cjf src-bzip.tar.bz2 src
tar tzf src-gzip.tar.gz
mkdir out-gz out-bz
tar xzf src-gzip.tar.gz -C out-gz
tar xjf src-bzip.tar.bz2 -C out-bz
ls -lR out-gz out-bz
gzip -k out-gz/src/a.txt
ls -l out-gz/src/a.txt*

Confirm the symlink survived packaging, both compression flavors unpack to matching trees, and standalone gzip produced a.txt.gz as expected.

Section Checkpoint

You should be able to create and extract .tar.gz and .tar.bz2 archives, list members before unpacking, compress and decompress individual files with gzip and bzip2, choose tar when directories are involved, and avoid path and overwrite mistakes during restore. Combined with solid cp/mv/ln skills, you can stage, package, and reassemble any file tree the exam throws at you.