Linux File Compression & Backup: Complete Guide to tar, gzip, bzip2, xz, zip, rsync, and scp

Introduction

The author shares a common ops nightmare: a missing backup after a power outage, and then demonstrates how a few command‑line tools can prevent data loss. The article covers the most useful Linux utilities for packing, compressing, and moving files.

tar: The Universal Archiver

tar creates archive files; it does not compress by default but can be combined with compression flags. A typical backup command is: tar -czvf backup_$(date +%Y%m%d).tar.gz /var/www/html/ Explanation of options: -c: create a new archive -z: compress with gzip (use -j for bzip2, -J for xz) -v: verbose output -f: specify the output file name (must be the last flag before the name)

To extract, replace -c with -x. Adding -C /target/path extracts to a specific directory. Mastering this command handles about 80 % of everyday archiving needs.

gzip vs bzip2 vs xz: Compression Rate Showdown

The article compares three common compressors: gzip: fastest, lower compression; ideal for log rotation with low CPU usage. bzip2: 10‑15 % smaller output than gzip but roughly twice as slow; often used on older systems. xz: best compression (about half the size of gzip) but high CPU and memory consumption; suitable when disk space is scarce or large images need transfer.

All three commands replace the original file (e.g., gzip file.txt creates file.txt.gz and removes file.txt). To keep the source, add -k or use a pipeline, for example:

tar cf - dir/ | xz > backup.tar.xz

zip: Cross‑Platform Compatibility

Although less common on Linux, zip is universally supported on Windows, macOS, and mobile devices. A typical command to create a high‑compression, recursive archive and optionally encrypt it is:

zip -r -9 secret_backup.zip ./conf/ && zip -e secure.zip sensitive_data.txt

Key options: -r: recurse into directories -9: maximum compression -e: prompt for encryption password

Extraction is done with unzip file.zip.

cpio & pax: Niche Archivers

These older tools are useful in specific scenarios. cpio works with standard input/output streams and is still used for kernel module installation and initramfs creation, e.g.:

find . -name "*.log" | cpio -ov > logs.cpio

pax

acts as a portable archiver that can read both tar and cpio formats, helpful when migrating data from legacy AIX or HP‑UX systems or when tar reports "invalid header" errors.

Remote Transfer & Synchronization: scp vs rsync

After compressing files, they need to be moved to a backup server. scp (Secure Copy) uses SSH for straightforward file transfer. Example: scp -P 2222 /opt/release.tar.gz [email protected]:/data/backups/ Adding -r copies directories and -C enables compression. However, scp restarts the transfer from the beginning after a disconnection, which is problematic for multi‑gigabyte files. rsync performs incremental synchronization, sending only changed parts. A typical command:

rsync -avz --progress -e "ssh -p 22" /data/web/ root@bak:/data/web_mirror/

Options explained: -a: archive mode (preserves permissions, timestamps, etc.) --delete: delete files on the destination that no longer exist on the source --partial: keep partially transferred files to allow resume

After an initial full sync, a subsequent run that only changes a small file like config.ini completes in seconds, even for source directories of 100 GB.

Conclusion: Practical Tips for Reliable Backups

Key take‑aways:

Daily backups: tar -cjf or tar -cJf to balance speed and size.

Cross‑team file sharing: use zip for universal compatibility.

Scheduled backups: combine rsync -a with cron for automated mirroring.

Large file transfers: avoid scp; prefer rsync --partial or run the command inside tmux for resilience.

Core mantra: "Pack with tar, choose compression based on scenario, share with zip, sync incrementally with rsync, and use the right parameters for resume‑able transfers."