Round-robin (default) in Nginx - Commands & Configuration

Introduction

When you have multiple servers to handle web traffic, you want to share the work evenly. Round-robin is a simple way to send each new visitor to the next server in line. This helps keep your website fast and reliable by not overloading one server.

When you want to balance web traffic evenly across several servers to avoid overload.

When you have multiple identical backend servers serving the same content.

When you want a simple, automatic way to distribute user requests without complex rules.

When you want to improve website availability by spreading requests in a cycle.

When you want to add or remove servers easily without changing the load balancing method.

Config File - nginx.conf

nginx.conf

http {
    upstream myapp {
        server backend1.example.com;
        server backend2.example.com;
        server backend3.example.com;
    }

    server {
        listen 80;

        location / {
            proxy_pass http://myapp;
        }
    }
}

The upstream block defines a group of backend servers named myapp. Each server line adds one backend server to the group.

The server block listens on port 80 and forwards all requests to the myapp group using proxy_pass.

By default, Nginx uses round-robin to send each new request to the next server in the list.

Commands

Check the Nginx configuration file for syntax errors before applying changes.

Terminal

nginx -t

Expected Output

nginx: the configuration file /etc/nginx/nginx.conf syntax is ok

nginx: configuration file /etc/nginx/nginx.conf test is successful

Reload Nginx to apply the new configuration without stopping the service.

Terminal

systemctl reload nginx

Expected Output

No output (command runs silently)

Send a request to Nginx to verify it forwards requests to backend servers using round-robin.

Terminal

curl -I http://localhost

Expected Output

HTTP/1.1 200 OK

Server: nginx

Date: Wed, 05 Jun 2024 12:00:00 GMT

Content-Type: text/html

Content-Length: 612

Connection: keep-alive

Key Concept

If you remember nothing else from this pattern, remember: round-robin sends each new request to the next server in order, balancing load evenly by default.

Common Mistakes

Forgetting to reload Nginx after changing the configuration.

Nginx keeps using the old settings until reloaded, so changes have no effect.

Always run 'systemctl reload nginx' or 'nginx -s reload' after editing the config.

Not checking configuration syntax before reloading.

Syntax errors can cause Nginx to fail to reload, leading to downtime.

Run 'nginx -t' to verify the config is valid before reloading.

Listing backend servers with wrong hostnames or IPs.

Nginx cannot forward requests if backend servers are unreachable.

Use correct and reachable server addresses in the upstream block.

Summary

Define backend servers in an upstream block to group them for load balancing.

Use proxy_pass in a server block to forward requests to the upstream group.

Reload Nginx after configuration changes to apply the new load balancing setup.

Practice

1. What does the default round-robin method in nginx do when balancing requests?

easy

A. It sends requests based on server load automatically.

B. It sends all requests to the first server until it fails.

C. It sends requests randomly to any server.

D. It sends each new request to the next server in the list, cycling through them.

Solution

1. Step 1: Understand round-robin behavior in nginx

   Round-robin means distributing requests evenly by cycling through servers one by one.

2. Step 2: Match behavior to options

   Only It sends each new request to the next server in the list, cycling through them. describes cycling through servers for each new request.

3. Final Answer:

   It sends each new request to the next server in the list, cycling through them. -> Option D

4. Quick Check:

   Round-robin = cycling requests evenly [OK]

Hint: Round-robin cycles through servers one by one [OK]

Common Mistakes:

• Thinking round-robin sends all requests to one server
• Confusing round-robin with random or load-based methods
• Assuming round-robin skips servers

2. Which of the following is the correct syntax to define a round-robin upstream block in nginx?

easy

A. upstream backend { round_robin; server server1; server server2; }

B. upstream backend { server server1.example.com; server server2.example.com; }

C. upstream backend { server server1; server server2; load_balance round_robin; }

D. upstream backend { server server1; server server2; balance round_robin; }

Solution

1. Step 1: Recall nginx upstream syntax

   The default round-robin method requires only listing servers inside an upstream block without extra keywords.

2. Step 2: Check each option for correctness

   upstream backend { server server1.example.com; server server2.example.com; } correctly lists servers simply. Options B, C, and D use invalid or unsupported directives.

3. Final Answer:

   upstream backend { server server1.example.com; server server2.example.com; } -> Option B

4. Quick Check:

   Default round-robin needs only server list [OK]

Hint: Round-robin needs only server list, no extra keywords [OK]

Common Mistakes:

• Adding unsupported directives like round_robin or balance
• Using incorrect keywords inside upstream block
• Forgetting semicolons after server lines

3. Given this nginx upstream configuration:

upstream backend {
  server server1.example.com;
  server server2.example.com;
  server server3.example.com;
}

What server will handle the 5th incoming request?

medium

A. server2.example.com

B. server1.example.com

C. server3.example.com

D. Depends on server load

Solution

1. Step 1: Understand round-robin cycling through servers

   Requests are sent in order: 1->server1, 2->server2, 3->server3, then cycle repeats.

2. Step 2: Calculate 5th request server

   Requests 4 and 5 cycle back: 4->server1, 5->server2.

3. Final Answer:

   server2.example.com -> Option A

4. Quick Check:

   5 mod 3 = 2, so server2 handles 5th request [OK]

Hint: Count requests modulo server count to find server [OK]

Common Mistakes:

• Assuming 5th request goes to server3
• Thinking load affects round-robin by default
• Starting count from zero instead of one

4. You configured nginx upstream with:

upstream backend {
  server server1.example.com;
  server server2.example.com;
  server server3.example.com down;
}

But requests are still sent to server3.example.com. What is the likely problem?

medium

A. The 'down' parameter is misspelled or misplaced.

B. Nginx does not support marking servers as down.

C. The 'down' parameter requires a reload, not just restart.

D. The 'down' parameter disables the server only if it is unreachable.

Solution

1. Step 1: Check correct usage of 'down' parameter

   The 'down' flag must be placed immediately after the server address without extra spaces or errors.

2. Step 2: Identify common syntax mistakes

   If 'down' is misspelled or misplaced, nginx ignores it and still sends requests to that server.

3. Step 3: Understand reload vs restart

   Nginx requires a reload (nginx -s reload) to apply config changes like 'down'; a restart may not apply changes immediately.

4. Final Answer:

   The 'down' parameter requires a reload, not just restart. -> Option C

5. Quick Check:

   Reload nginx to apply 'down' flag changes [OK]

Hint: Reload nginx after config changes to apply 'down' flag [OK]

Common Mistakes:

• Misspelling 'down' as 'dwon' or 'dnw'
• Placing 'down' on a separate line
• Not reloading nginx after config change

5. You want to ensure nginx sends requests evenly to three servers but skip one server temporarily without removing it from config. Which is the best way to do this using round-robin?

hard

A. Mark the server as 'down' in the upstream block to skip it.

B. Remove the server line from the upstream block temporarily.

C. Add 'weight=0' to the server to disable it.

D. Use 'max_fails=0' to prevent requests to that server.

Solution

1. Step 1: Understand how to skip a server temporarily

   Marking a server as 'down' tells nginx to skip it without removing it from config.

2. Step 2: Evaluate other options

   Removing the server line is permanent. 'weight=0' is invalid syntax. 'max_fails=0' disables failure detection but not requests.

3. Final Answer:

   Mark the server as 'down' in the upstream block to skip it. -> Option A

4. Quick Check:

   'down' flag skips server temporarily [OK]

Hint: Use 'down' flag to skip server without removing it [OK]

Common Mistakes:

• Removing server line instead of marking down
• Using invalid 'weight=0' syntax
• Confusing max_fails with disabling server