Apache Reverse Proxy Setup

Overview

This lecture explains how to configure and manage the Apache HTTP Server as a reverse proxy, focusing on setup, load balancing, failover, and health checks.

Introduction to Reverse Proxy

• Apache httpd can function as a reverse proxy, forwarding client requests to backend servers.
• Clients only see the reverse proxy as the source of content, not the backend infrastructure.
• Common reasons for using a reverse proxy include security, load balancing, high-availability, and centralized authentication.

Simple Reverse Proxying

• Use the ProxyPass directive to map incoming requests to a backend server or cluster.
• The ProxyPassReverse directive rewrites headers so responses appear from the reverse proxy, not the backend.
• Specific paths (e.g., /images) can be proxied while other requests are handled locally.

Clusters and Balancers

• A "balancer" is a group of backend servers (BalancerMembers) managed by Apache for load balancing.
• Define a balancer with <Proxy> and BalancerMember directives and specify load balancing methods with ProxySet.

Balancer and BalancerMember Configuration

• Balancer settings can be customized; e.g., loadfactor adjusts traffic distribution, and timeout sets response limits.
• Example: Assign higher traffic to a server with loadfactor=3.

Failover

• Configure hot spares and hot standbys to ensure continued availability if backend servers fail.
• Workers in drain, stopped, or error state are replaced by hot spares; standbys are used if all others fail.
• Load balancer sets are tried in order, including backup sets if necessary.

Balancer Manager

• The balancer-manager tool provides a web interface to display and modify balancer and worker settings in real time.
• Enable with <Location "/balancer-manager"> and restrict access for security.
• Use BalancerPersist to make changes persistent after restarts.

Dynamic Health Checks

• The mod_proxy_hcheck module can check the health of workers dynamically and out-of-band.
• Workers can be "pinged" before sending requests to verify availability.

BalancerMember Status Flags

• Workers have various statuses: Ok (available), Dis (disabled), Stop (administratively stopped), Ign (ignore-errors), Spar (hot spare), Stby (hot standby), Err (error), Drn (drain mode), HcFl (health check failed).

Key Terms & Definitions

• Reverse Proxy — A server that forwards client requests to backend servers and returns their responses.
• ProxyPass — Directive to map client requests to backend servers.
• ProxyPassReverse — Directive to rewrite backend response headers.
• Balancer — A group of backend servers managed for load balancing.
• BalancerMember — An individual server in a balancer group.
• loadfactor — Parameter controlling relative traffic to each worker.
• Balancer Manager — Web interface for managing balancer settings.
• Hot Spare — A backup server that replaces failed workers in real time.
• Hot Standby — A server activated only if no other workers or spares are available.

Action Items / Next Steps

• Review and experiment with ProxyPass, ProxyPassReverse, and <Proxy> configuration in your Apache setup.
• Secure any balancer-manager URLs before enabling in production.
• Consider enabling mod_proxy_hcheck for advanced health checks.
• Read more on failover configuration and status flags for robust proxy setups.