Audit logging in Hadoop

Introduction

Audit logging helps you keep track of who did what and when in your Hadoop system. It is useful for security and troubleshooting.

You want to see who accessed or changed files in HDFS.

You need to check if someone tried to do something they shouldn't.

You want to keep a record of user actions for compliance rules.

You are troubleshooting problems and want to know the sequence of events.

You want to monitor system usage patterns over time.

Syntax

Hadoop

hadoop.security.authorization=true
hadoop.security.audit.logger=org.apache.hadoop.security.authorize.AuditLogger
hadoop.security.audit.loggers=org.apache.hadoop.security.authorize.AuditLogger

These settings go into the core-site.xml or hdfs-site.xml configuration files.

Make sure to restart Hadoop services after changing these settings.

Examples

This enables security authorization in Hadoop, which is needed for audit logging.

Hadoop

<property>
  <name>hadoop.security.authorization</name>
  <value>true</value>
</property>

This sets the class that handles audit logging.

Hadoop

<property>
  <name>hadoop.security.audit.logger</name>
  <value>org.apache.hadoop.security.authorize.AuditLogger</value>
</property>

This configures the audit loggers to use.

Hadoop

<property>
  <name>hadoop.security.audit.loggers</name>
  <value>org.apache.hadoop.security.authorize.AuditLogger</value>
</property>

Sample Program

This is a complete example of the audit logging configuration in an XML format for Hadoop's core-site.xml or hdfs-site.xml file.

Hadoop

<configuration>
  <property>
    <name>hadoop.security.authorization</name>
    <value>true</value>
  </property>
  <property>
    <name>hadoop.security.audit.logger</name>
    <value>org.apache.hadoop.security.authorize.AuditLogger</value>
  </property>
  <property>
    <name>hadoop.security.audit.loggers</name>
    <value>org.apache.hadoop.security.authorize.AuditLogger</value>
  </property>
</configuration>

Important Notes

Audit logs are usually stored in the Hadoop log directory, check your log4j or logback configuration for exact location.

Audit logging can generate a lot of data, so monitor disk space and log rotation settings.

Ensure your Hadoop cluster has proper permissions so audit logs cannot be tampered with.

Summary

Audit logging tracks user actions in Hadoop for security and troubleshooting.

Enable it by setting hadoop.security.authorization to true and configuring audit logger classes.

Audit logs help you see who accessed or changed data and when.

Practice

1. What is the main purpose of audit logging in Hadoop?

easy

A. To backup data automatically

B. To improve data processing speed

C. To compress files for storage

D. To track user actions for security and troubleshooting

Solution

1. Step 1: Understand audit logging function

   Audit logging records user actions like access and changes in Hadoop.

2. Step 2: Identify main goal

   This helps with security checks and troubleshooting issues.

3. Final Answer:

   To track user actions for security and troubleshooting -> Option D

4. Quick Check:

   Audit logging = Track user actions [OK]

Hint: Audit logs = user actions tracking for security [OK]

Common Mistakes:

• Confusing audit logging with data backup
• Thinking audit logs speed up processing
• Mixing audit logging with file compression

2. Which configuration property must be set to true to enable audit logging in Hadoop?

easy

A. hadoop.logging.audit.enabled

B. hadoop.security.audit.enable

C. hadoop.audit.logging.active

D. hadoop.security.authorization

Solution

1. Step 1: Identify correct property name

   The official property to enable audit logging is hadoop.security.audit.enable.

2. Step 2: Confirm value to enable

   Setting it to true activates audit logging features.

3. Final Answer:

   hadoop.security.audit.enable -> Option B

4. Quick Check:

   Enable audit logging = hadoop.security.audit.enable true [OK]

Hint: Enable audit with hadoop.security.audit.enable true [OK]

Common Mistakes:

• Using incorrect property names
• Setting property to false
• Confusing audit logging with general logging

3. Given the following audit log entry:

2024-06-01 10:15:30 INFO User: alice accessed /data/input/file1.txt

What does this log entry indicate?

medium

A. User alice modified the file /data/input/file1.txt

B. User alice deleted the file /data/input/file1.txt

C. User alice accessed the file /data/input/file1.txt

D. User alice created the file /data/input/file1.txt

Solution

1. Step 1: Read the log message carefully

   The log states "User: alice accessed /data/input/file1.txt" which means alice opened or read the file.

2. Step 2: Understand the action word

   "accessed" means viewing or reading, not modifying or deleting.

3. Final Answer:

   User alice accessed the file /data/input/file1.txt -> Option C

4. Quick Check:

   Access means read/view, not modify [OK]

Hint: Look for action verb in log entry to know user activity [OK]

Common Mistakes:

• Assuming access means modification
• Confusing access with deletion
• Ignoring the exact log message wording

4. You enabled audit logging by setting hadoop.security.authorization to true, but no audit logs are generated. What is the most likely cause?

medium

A. Audit logger classes are not configured properly

B. Hadoop service is not running

C. User permissions are too high

D. Audit logging is disabled by default

Solution

1. Step 1: Check audit logging prerequisites

   Besides enabling authorization, audit logger classes must be configured to capture logs.

2. Step 2: Identify missing configuration

   If audit logger classes are missing or misconfigured, no logs will appear despite enabling authorization.

3. Final Answer:

   Audit logger classes are not configured properly -> Option A

4. Quick Check:

   Enable + configure logger classes = audit logs [OK]

Hint: Enable plus configure logger classes to get audit logs [OK]

Common Mistakes:

• Assuming service not running causes no logs
• Thinking user permissions block audit logs
• Ignoring need to configure logger classes

5. You want to audit all user actions on Hadoop files and see who changed data and when. Which combination of steps is correct?

hard

A. Set hadoop.security.authorization=true and configure audit logger classes

B. Set hadoop.security.authorization=false and enable debug logging

C. Only configure audit logger classes without enabling authorization

D. Enable file system replication and audit logging

Solution

1. Step 1: Enable security authorization

   Setting hadoop.security.authorization=true activates audit logging capability.

2. Step 2: Configure audit logger classes

   These classes capture and record user actions in logs.

3. Final Answer:

   Set hadoop.security.authorization=true and configure audit logger classes -> Option A

4. Quick Check:

   Enable authorization + logger classes = full audit [OK]

Hint: Enable authorization and logger classes for full audit [OK]

Common Mistakes:

• Disabling authorization but expecting audit logs
• Relying on debug logging for audit
• Confusing replication with audit logging