gzip Command - Comprehensive Reference Guide

Overview

Purpose

The gzip command compresses files using the DEFLATE compression algorithm. It reduces file size significantly, making it ideal for storage and network transfer. Most commonly used for single-file compression; combine with tar for multiple files.

Basic Syntax

gzip [options] filename

Key Concept

Gzip replaces the original file with a compressed version (.gz extension). Use gunzip or gzip -d to decompress. For archives with multiple files, use tar with gzip: tar -czf archive.tar.gz files/

Compression Concepts

Compression Levels

-1 (fastest, worst compression) through -9 (slowest, best compression). Default is -6. Higher compression (-9) takes more CPU time but creates smaller files.

File Replacement

By default, gzip replaces the original file with the compressed version. Use -c to write to stdout and keep the original, or -k to keep the original file.

Compression Level	Speed	Ratio	Use Case
-1 (fastest)	Very fast	Poor	Real-time compression
-6 (default)	Balanced	Good	Most common
-9 (best)	Slow	Best	Archival, storage

Common Options

Option	Description
`-c`	Write to stdout, don't modify original
`-d`	Decompress (same as gunzip)
`-k`	Keep original file (don't delete)
`-v`	Verbose - show compression details
`-r`	Recursively compress directory
`-t`	Test compressed file integrity
`-#`	Compression level (1-9)

10 Detailed Examples

Basic File Compression

Before

ls -lh myfile.txt
-rw-r--r-- 1 user user 50M myfile.txt

Compress file

gzip myfile.txt

After

ls -lh myfile.txt.gz
-rw-r--r-- 1 user user 5M myfile.txt.gz

What This Does:

Compresses myfile.txt and replaces it with myfile.txt.gz. The original file no longer exists. File size reduced from 50MB to 5MB (90% compression on this text data).

Common Usage:

Perfect for log files, backups, and data archival. Compression ratio varies by file type (text compresses well, already-compressed files like PNG don't compress much further).

Decompress File

Method 1: gunzip

gunzip myfile.txt.gz

Method 2: gzip -d

gzip -d myfile.txt.gz

[file is decompressed back to myfile.txt]

What This Does:

Decompresses the .gz file back to original. Both gunzip and gzip -d do the same thing. The original .gz file is deleted.

Quick Tip:

gunzip is shorter to type and is the standard decompression tool. Most systems have it as an alias or separate command.

Keep Original File with -k

Compress but keep original

gzip -k large_dataset.csv

Check files

ls -lh large_dataset.*

-rw-r--r-- 1 user user 100M large_dataset.csv
-rw-r--r-- 1 user user 10M large_dataset.csv.gz

What This Does:

The -k flag keeps the original file after compression. Useful when you need both compressed and uncompressed versions for compatibility.

Use Case:

Archive original files while providing compressed versions for download/transfer. Backup systems often do this.

Compress to stdout with -c

Pipe compressed output

gzip -c myfile.txt | tee backup.gz | wc -c

5242880

Compress and send via SSH

gzip -c large_file.log | ssh remote.host 'cat > /backups/file.log.gz'

What This Does:

The -c flag writes compressed output to stdout instead of creating a .gz file. Allows piping for efficient network transfers or further processing.

Powerful Patterns:

Compress and pipe to remote system
Compress while monitoring size
Compress for direct stdout inspection

Compression Levels: Speed vs Ratio

Fastest compression (-1)

time gzip -1 -c large_file.txt > output1.gz

real 0m1.234s

Default compression (-6)

time gzip -6 -c large_file.txt > output6.gz

real 0m3.567s

Best compression (-9)

time gzip -9 -c large_file.txt > output9.gz

real 0m8.234s

Compare file sizes

ls -lh output*.gz

-rw-r--r-- 1 user user 12M output1.gz
-rw-r--r-- 1 user user 10M output6.gz
-rw-r--r-- 1 user user 9.8M output9.gz

What This Shows:

Trade-off between speed and compression ratio. -1 is 6.7x faster than -9 but produces 22% larger file. Choose based on your priorities.

Recommendations:

-1 - Real-time logging, streaming
-6 - General purpose (default, usually best)
-9 - Archival, storage (minimal space premium)

Verbose Output with -v

Compress with verbose info

gzip -v *.log

app.log: 83.2% (100M -> 16.8M)
system.log: 72.5% (50M -> 13.7M)
access.log: 68.9% (75M -> 23.4M)

What This Does:

Shows compression details for each file: percentage saved, original size, and compressed size. Useful for monitoring batch operations.

Batch Operations:

Combine with * wildcard to compress multiple files at once. Each file gets its own .gz file.

Recursive Directory Compression

Compress all files in directory tree

gzip -r /var/log/

Verify

find /var/log -name "*.gz" | head

/var/log/apache2/access.log.gz
/var/log/apache2/error.log.gz
/var/log/apt/history.log.gz
/var/log/auth.log.gz
/var/log/cron.log.gz

What This Does:

The -r flag recursively compresses all files in directory and subdirectories. Each file individually becomes .gz; directories aren't changed.

Important Note:

-r doesn't create a single archive like tar. It compresses each file individually. For a single compressed archive, use tar: tar -czf archive.tar.gz directory/

Test Compressed File Integrity

Test file for corruption

gzip -t backup.tar.gz

[no output = file is OK]

If corrupted:

gzip: backup.tar.gz: invalid compressed data--crc error

What This Does:

Tests the integrity of compressed file without decompressing. Returns success silently or shows error. Critical for verifying backups.

Best Practice:

Always test backups after compression: gzip -t backup.gz && echo "Backup OK" || echo "Backup corrupted"

tar with gzip Integration (tar.gz)

Create compressed tar archive

tar -czf backup.tar.gz /home/user/documents/

Extract (gzip handles automatically)

tar -xzf backup.tar.gz

List contents without extracting

tar -tzf backup.tar.gz | head

home/user/documents/file1.txt
home/user/documents/file2.doc
home/user/documents/subdir/file3.pdf

What This Does:

Tar creates archives of multiple files/directories; gzip compresses the result. The combination -z flag automatically handles gzip. Most efficient for backing up directories.

Production Backup Pattern:

tar -czf /backups/daily-$(date +%Y%m%d).tar.gz /critical/data/ creates timestamped compressed backups.

Advanced: Piping and Network Transfer

Backup directly to remote system

tar -c /important/data | gzip -9 | ssh backup@archive.local 'cat > /backups/data-$(date +%Y%m%d).tar.gz'

Remote verification

ssh backup@archive.local 'gzip -t /backups/data-*.tar.gz && echo "All backups valid"'

What This Does:

Demonstrates advanced pattern: tar archives, gzip compresses with level 9, piped over SSH to remote system. Creates timestamped backup without local disk overhead.

Breaking Down the Command:

tar -c - Create archive, output to stdout
gzip -9 - Compress with best ratio
| ssh - Pipe to remote system
cat > /path/file - Write to remote file
$(date +%Y%m%d) - Add timestamp

Production Enterprise Backup:

This pattern scales to terabytes; compressed stream goes directly to network without temporary files.

Comparison: gzip vs Other Compression Tools

Tool	Compression	Speed	Ratio	Best For
gzip	DEFLATE	Fast	Good	General purpose (most common)
bzip2	Burrows-Wheeler	Slower	Better	Maximum compression
xz	LZMA2	Slowest	Best	Extreme compression (Linux kernels)
zip	DEFLATE	Fast	Good	Cross-platform archives

Practical System Administration Scenarios

Log Rotation and Compression

Logrotate typically compresses old logs: compress postrotate /usr/lib/rsyslog/rsyslog-rotate.sh applies gzip to rotated logs automatically.

Backup Strategy

Daily backup: tar -czf /backups/daily-$(date +%Y%m%d).tar.gz /important/data. Weekly: gzip -9 for maximum compression. Test with gzip -t.

Distribution Packages

Most Linux source distributions use .tar.gz format. Extract with: tar -xzf package.tar.gz

Troubleshooting

Problem: File not compressing well

Already-compressed formats (PNG, JPEG, MP4) compress poorly. Text, logs, and data compress best. Check compression ratio with gzip -v.

Problem: Permission denied when decompressing

Check file permissions and target directory: ls -l file.gz. Ensure write access to output directory.

Problem: "unexpected end of file" error

File is corrupted. Test with gzip -t first. Transfer over network may have corrupted the file. Re-download and verify.