If you work with Linux, you will eventually spend time reading logs, searching configuration files, and cleaning messy text output.

At first, many people try to do this manually.

Open file. Scroll. Search. Copy text. Repeat.

That works for small tasks.

But sysadmins do not work that way.

They use command-line tools that solve problems in seconds.

Three of the most useful tools are:

• grep
• awk
• sed

If you learn these well, your Linux troubleshooting speed improves a lot.

This post shows practical examples, not textbook definitions.

Why These Three Commands Matter
Imagine these real situations:

You want to find failed SSH login attempts.

You want to extract IP addresses from logs.

You want to replace a wrong server name in a config file.

You want to count repeated entries.

Doing this manually wastes time.

This is where these commands help.

Think of them like this:

• grep = find text
• awk = extract and process columns
• sed = edit and transform text

1. grep: Find What Matters Fast
grep searches for matching text.

Basic syntax:

grep "pattern" filename

Example:

grep "error" app.log

Output:

database connection error
api timeout error

This finds lines containing the word error.

Useful grep Options

Ignore case

grep -i "error" app.log

Matches:

• Error
• ERROR
• error

Helpful when log formats are inconsistent.

Show line numbers

grep -n "server" nginx.conf

Output:

12:server_name example.com;
45:server_tokens off;

Good for config debugging.

Invert match

Show lines that do NOT match:

grep -v "INFO" app.log

Useful when removing noisy logs.

Recursive search

Search inside directories:

grep -r "Listen 80" /etc/apache2

Very useful for config hunting.

Real SysAdmin Example

Find failed SSH login attempts:

grep "Failed password" /var/log/auth.log

Sample output:

Failed password for root from 192.168.1.10
Failed password for admin from 10.0.0.5

This quickly shows suspicious login attempts.

2. awk: Extract and Process Data

awk is excellent when data has columns.

Example file:

alice 5000 IT
bob 7000 HR
john 6500 DevOps

Print first column:

awk '{print $1}' employees.txt

Output:

alice
bob
john

Explanation:

• $1 = first column
• $2 = second column
• $3 = third column

Print Multiple Columns

awk '{print $1, $3}' employees.txt

Output:

alice IT
bob HR
john DevOps

Filter by Condition

Show salaries above 6000:

awk '$2 > 6000 {print $1, $2}' employees.txt

Output:

bob 7000
john 6500

This is very useful for reports.

Real SysAdmin Example

Check logged-in users:

who

Example output:

pawan pts/0 2026-05-18 10:30
john pts/1 2026-05-18 11:00

Extract usernames:

who | awk '{print $1}'

Output:

pawan
john

3. sed: Stream Editing Made Simple

sed helps modify text.

Basic replacement:

sed 's/old/new/' file.txt

Example:

sed 's/dev/prod/' config.txt

If file contains:

server=dev

Output:

server=prod

Replace All Matches

Without global flag, only first match changes.

Use:

sed 's/error/warning/g' app.log

g = global replacement

Delete Lines

Delete blank lines:

sed '/^$/d' file.txt

Very useful when cleaning files.

Edit File Directly

sed -i 's/localhost/db-server/' config.ini

Be careful.

This changes the actual file.

Real SysAdmin Example

Update nginx config:

Before:

server_name oldsite.com;

Command:

sed -i 's/oldsite.com/newsite.com/' nginx.conf

Fast and practical.

Combining Commands Like a SysAdmin

The real power comes from combining tools.

Example:

Find failed SSH attempts and extract IP addresses:

grep "Failed password" /var/log/auth.log | awk '{print $11}'

Output:

192.168.1.10
10.0.0.5

Count repeated IPs:

grep "Failed password" /var/log/auth.log | awk '{print $11}' | sort | uniq -c

Sample output:

5 192.168.1.10
2 10.0.0.5

Now you know which IP is attacking most.

This is real troubleshooting.

Common Beginner Mistakes

Using grep for everything

Yes, grep is useful.

But if you need column processing, use awk.

Editing files with sed without backup

This:

sed -i 's/test/prod/' file.conf

changes the file immediately.

Safer:

cp file.conf file.conf.bak

then edit.

Forgetting quotes

Wrong:

grep error file.txt

Better:

grep "error" file.txt

Especially for complex patterns.

Quick Comparison