Hosting a website on EC2: web servers, Apache httpd, /var/www/html, port 80 and curl

By Ravindra Bagale, trainer · LinkedIn · Last updated: 8 October 2026

1. Why hosting: from a laptop to a live website

Why. Developers write a website on their laptops. Nobody in the world can open it there. Someone has to put it on a server that runs day and night and that anyone can reach over the internet. That someone is the cloud engineer, and putting it there is called hosting.

What. Software usually moves through three places:

1. Development: developers write the code on their own laptops.
2. Test: testers check it on a test server.
3. Production: the live server that real users open.

Cloud engineers don't write the website. They build and run the servers it lives on. There are two kinds of hosting:

1. Website hosting: the server sends ready web pages (HTML, CSS, JavaScript, images).
2. API (backend) hosting: the server runs backend code and sends data, usually JSON.

This chapter does website hosting from start to finish on one EC2 server.

2. What happens when someone opens a website

Why. If you know the path a request takes, you know where to look when a site doesn't open.

What. Suppose your site is www.mysite.com and someone opens http://www.mysite.com/about.html:

1. The browser asks DNS for the IP address of www.mysite.com. DNS answers with the server's public IP.
2. The browser sends the request to that IP. It uses http and gives no port number, so the port is 80, the default for HTTP. For https the default is 443.
3. On the server, a program is listening on port 80, 24 hours a day. That program is the web server.
4. The web server reads the request ("give me about.html") and looks for that file in its web folder. On Amazon Linux with Apache, that folder is /var/www/html.
5. If the file is there, it goes back to the browser, which shows it. If not, the answer is 404 Not Found.
6. If the URL has no file name at all, like http://www.mysite.com/, the web server sends the folder's home page, index.html.

You learned ports 80 and 443 in Public vs private IP, ports and IP classes. Domain names and DNS get their own chapter later; today we use the server's public IP directly.

Example 1: three URLs on the same server.

1. http://PUBLIC_IP/ gives /var/www/html/index.html.
2. http://PUBLIC_IP/about.html gives /var/www/html/about.html.
3. http://PUBLIC_IP/contact.html, when there is no such file, gives 404 Not Found.

3. What a web server is

Why. A server machine runs many programs. Only one kind of program can answer a browser: one that listens on port 80 (or 443), understands an HTTP request and sends back a response. That program is a web server.

What. The common ones:

1. Apache HTTP Server: open source and free. On Amazon Linux the package and the program are called httpd.
2. Nginx: also open source and free. A company behind it sells paid support and extra features.
3. IIS: Microsoft's web server on Windows, used a lot for .NET sites.
4. Lighttpd: a small, light web server.

Two more names you will hear:

1. Gunicorn and Uvicorn are app servers. They run Python backend code, usually behind Nginx or Apache.
2. XAMPP and WAMP are bundles for practice on a laptop. One installer gives you Apache, MySQL or MariaDB, and PHP.

In this chapter we use Apache. Nginx follows later; the idea is the same.

4. Website hosting vs API hosting

Why. "Hosting" means two different jobs, and they need different things installed.

What. Suppose we are Instagram. People open it on phones of many sizes, tablets, laptops and even car screens. If the server sent one ready HTML page with a fixed layout, it would look wrong on many of those screens.

So such apps split the work:

1. The app on your phone asks the server for data only, for example the profile of user number 1001.
2. On the server, backend code (Python, Node.js, Java, .NET) reads that data from a database (MySQL, MongoDB).
3. The server sends the data back as JSON, a simple key-value format.
4. The app knows its own screen size and draws the screen itself.

Example 2: a JSON answer.

{"username": "ravi", "followers": 500}

Text values are in double quotes; numbers are not.

So:

1. Website hosting: the server sends whole pages. You need a web server and the files.
2. API (backend) hosting: the server sends data. You need the language (Python, Node.js, Java or .NET), an app server, often a database, and a web server in front. Developers write that code; the cloud engineer sets up the server and deploys it.

5. Static vs dynamic websites

Why. Before you host a site you must know which kind it is, because a dynamic site needs much more on the server.

What. A static website is made only of files that the browser understands:

1. HTML: the structure and content of the page.
2. CSS: how it looks (colours, fonts, layout).
3. JavaScript: code that runs in the browser, not on the server.
4. Images and videos.

Every visitor gets the same files. The server only finds the file and sends it.

A dynamic website also needs:

1. A server-side language, such as Python, PHP, Node.js or Java.
2. A database.

The server builds the page for each visitor from their data. Suppose a reels app: each user sees different reels, chosen from what they watched before, which is stored in a database.

Example 3: HTML doesn't calculate. Put this in an HTML file and open it:

a=2
b=3
c=a+b

The browser shows exactly that text. It doesn't print 5. HTML only describes what is on the page; to calculate you need JavaScript in the browser, or a language on the server.

Today we host a static site. Dynamic sites come in later chapters.

6. Step 1: launch an Amazon Linux server and connect

What. You did this in Root user, IAM users, MFA and your first EC2 server and Inside EC2:

1. In the EC2 console, launch an instance with the Amazon Linux 2023 AMI.
2. Use your key pair and a Security Group that allows SSH (port 22).
3. Connect with SSH as ec2-user.

"Live" means the site is on a server that anyone can reach over the internet. Your EC2 server with a public IP is exactly that. Ubuntu servers work a little differently (another package manager, and the package is called apache2); this chapter is for Amazon Linux.

7. Step 2: install Apache with yum

Why. A fresh server has no web server. Check first:

[ec2-user@ip-172-31-xx-xx ~]$ httpd -v
-bash: httpd: command not found

command not found means there is no program called httpd in the folders Linux searches for commands. In other words: not installed.

What. One command installs it:

sudo yum install httpd

Read it in four parts:

1. sudo: run as root. Installing software for the whole system is an admin job.
2. yum: the package manager, the tool that installs, updates and removes software.
3. install: the action. remove would uninstall.
4. httpd: the package name of Apache.

What the package manager does for you.

1. It knows where to download from. The download addresses (repositories) are set in .repo files under /etc/yum.repos.d/.
2. It knows what else httpd needs (its dependencies) and installs those first, in the right order.
3. It downloads the packages (compressed RPM files) and installs them.
4. It works from any folder; you don't need to cd anywhere first.

How it looks. Here is the start of the real output on Amazon Linux 2023 (long lines shortened with ...):

[ec2-user@ip-172-31-xx-xx ~]$ sudo yum install httpd
...
Dependencies resolved.
...
Installing:
 httpd                 x86_64    2.4.68-1.amzn2023.0.1    amazonlinux    ...
Installing dependencies:
 apr                   x86_64    ...
 apr-util              x86_64    ...
 httpd-core            x86_64    ...
 httpd-filesystem      noarch    ...
 httpd-tools           x86_64    ...
 ...
Installing weak dependencies:
 mod_http2             x86_64    ...
 mod_lua               x86_64    ...
 ...
Transaction Summary
...
Install  14 Packages

Total download size: ...
Installed size: ...
Is this ok [y/N]:

1. Installing: the package you asked for.
2. Installing dependencies: packages httpd cannot work without.
3. Installing weak dependencies: useful extras that come along by default.
4. Total download size is smaller than Installed size, because packages are downloaded compressed and unpacked on the server.
5. Is this ok [y/N]: asks you to confirm. The capital N is the default: just pressing Enter means no.

The version and the number of packages on your server may differ; they change as Amazon Linux gets updates.

Example 4: answering N, then installing with -y. If you type N, nothing is installed:

Is this ok [y/N]: N
Operation aborted.

Add -y to answer yes in advance. This is the command we use:

sudo yum install httpd -y

It downloads, installs and ends with:

Complete!

8. Step 3: check the install

Why. Before going further, confirm that Apache is really there.

How. Ask for its version:

[ec2-user@ip-172-31-xx-xx ~]$ httpd -v
Server version: Apache/2.4.68 (Amazon Linux)
Server built:   ...

A version means it is installed. command not found means it is not.

Version flags differ between programs.

1. Many programs accept --version.
2. Apache's httpd uses -v for the version.
3. httpd -V (capital V) prints the version plus how it was built.

9. Step 4: software vs service, start and status

Why. Installing is not enough. A web server must keep running and listening on port 80, day and night.

What. Some software you open, use and close, like an editor. Other software runs in the background all the time and waits for work. That is a service. Apache, Nginx and MySQL are services.

How. Start it, then check it:

[ec2-user@ip-172-31-xx-xx ~]$ sudo service httpd start
Redirecting to /bin/systemctl start httpd.service

The first line is not an error. On Amazon Linux 2023, service is a small helper that passes the job to systemctl, the tool that manages services, and tells you so. sudo service httpd start and sudo systemctl start httpd do the same thing.

Now the status:

[ec2-user@ip-172-31-xx-xx ~]$ sudo service httpd status
Redirecting to /bin/systemctl status httpd.service
● httpd.service - The Apache HTTP Server
     Loaded: loaded (/usr/lib/systemd/system/httpd.service; disabled; ...)
     Active: active (running) since ...
...
... systemd[1]: Started httpd.service - The Apache HTTP Server.
... httpd[...]: Server configured, listening on: port 80

Read three things:

1. Loaded: loaded: the service is installed and known.
2. Active: active (running) (green on your screen): it is running now.
3. listening on: port 80: it is waiting for browsers.

Example 5: the three answers you can get from status. When httpd is not installed at all:

Redirecting to /bin/systemctl status httpd.service
Unit httpd.service could not be found.

The other two:

1. Installed but not started: Active: inactive (dead). Fix: sudo service httpd start.
2. Running: Active: active (running).

Start at every boot. start runs Apache now, but after a reboot (or a stop and start of the instance) it stays off. A freshly installed httpd is not set to start at boot; that's the disabled in the Loaded: line. Turn that on once:

[ec2-user@ip-172-31-xx-xx ~]$ sudo systemctl enable httpd
Created symlink /etc/systemd/system/multi-user.target.wants/httpd.service → /usr/lib/systemd/system/httpd.service.

10. When Apache won't start

Why. If start fails, the status tells you why. Learn to read it instead of reinstalling.

What. Three common causes:

1. Not installed. Status says Unit httpd.service could not be found. Install it first.
2. Port 80 is already taken. Only one program can listen on a port. If another web server, say Nginx, already uses port 80, Apache can't.
3. A mistake in a config file. Not likely on a fresh install, but it happens after edits.

Example 6: port 80 already in use. The start fails, sudo service httpd status shows failed, and its last lines (or sudo journalctl -u httpd) contain:

(98)Address already in use: AH00072: make_sock: could not bind to address [::]:80
(98)Address already in use: AH00072: make_sock: could not bind to address 0.0.0.0:80
no listening sockets available, shutting down

Fix: stop the other program (for Nginx, sudo service nginx stop), then sudo service httpd start again.

11. Step 5: your first page in /var/www/html

Why. Before you put a developer's real code on the server, test the setup with one tiny page. If that page opens, the web server, the folder, the home page and the port all work. Then any later problem is in the code, not in the server.

What.

1. /var/www/html is Apache's web folder (its DocumentRoot). It is created when the httpd package is installed, and it is empty.
2. The home page must be called exactly index.html. Apache's setting DirectoryIndex index.html makes it the page for a URL without a file name.
3. Until you add index.html, Apache shows its own test page, titled "It works! Apache httpd".

Example 7: go to the folder. Type cd /var/ and press Tab to complete each part:

[ec2-user@ip-172-31-xx-xx ~]$ cd /var/www/html
[ec2-user@ip-172-31-xx-xx html]$ pwd
/var/www/html
[ec2-user@ip-172-31-xx-xx html]$ ls
[ec2-user@ip-172-31-xx-xx html]$

Empty.

Example 8: without sudo you can't save. The folder belongs to root:

[ec2-user@ip-172-31-xx-xx html]$ ls -ld /var/www/html
drwxr-xr-x. 2 root root 6 Jan  5 10:15 /var/www/html

rwxr-xr-x: only the owner, root, may write. You learned this in chmod in depth. So nano index.html as ec2-user opens the editor, but saving fails at the bottom of the screen:

[ Error writing index.html: Permission denied ]

touch index.html fails the same way:

touch: cannot touch 'index.html': Permission denied

Example 9: create it with sudo. Leave nano without saving (Ctrl+X, then N), then:

sudo nano index.html

Type one line:

<h1>Hello from my EC2 server</h1>

Save and exit: Ctrl+X, then Y, then Enter. Check:

[ec2-user@ip-172-31-xx-xx html]$ ls -l
total 4
-rw-r--r--. 1 root root 34 Jan  5 10:15 index.html

The file belongs to root, and everyone can read it (r-- for others). That's all Apache needs.

12. Step 6: open port 80 in the Security Group

Why. Apache is running and the page is there, but if you open http://PUBLIC_IP/ now, the browser just waits and then gives up. Something in front of the server blocks the request: the Security Group.

What. A Security Group is a firewall for your instance. It has two lists of rules:

1. Inbound rules: traffic that comes to your server. At launch there is usually only one: SSH on port 22. Anything without a rule is blocked.
2. Outbound rules: traffic your server starts towards the internet, for example yum downloading packages. By default all outbound traffic is allowed.

Replies to a request that was allowed in go back out automatically. You don't need an outbound rule for your website's answers.

How. In the EC2 console:

1. Open Instances and select your instance.
2. Open the Security tab.
3. Click the Security group link (its ID starts with sg-).
4. Click Edit inbound rules.
5. Click Add rule.
6. Type: choose HTTP. The port 80 fills in by itself.
7. Source: choose Anywhere-IPv4, which is 0.0.0.0/0, meaning "from any IP".
8. Click Save rules.

13. Step 7: open your website

How.

1. In the EC2 console, copy your instance's Public IPv4 address.
2. In the browser, type http:// and then the IP: http://PUBLIC_IP/.
3. You see Hello from my EC2 server.

Example 10: proofs that it's really live.

1. Open the same address on your phone, on mobile data. It works there too.
2. In the browser, right-click the page and choose View Page Source. You see exactly the line you typed in nano.
3. Back on the server, history lists every command you used today. Keep it as your notes.

14. Debugging with curl

Why. When the browser shows nothing, you need to know: is the problem inside the server, or outside it? curl answers that in one command.

What. curl is a command-line program that asks a web server for a page, the same kind of request a browser sends, but it prints the raw HTML instead of drawing it. localhost means "this same machine", so the request never passes the Security Group.

Example 11: on the server.

[ec2-user@ip-172-31-xx-xx ~]$ curl http://localhost
<h1>Hello from my EC2 server</h1>

1. You see your HTML: Apache and the file are fine. The problem is outside: the Security Group's HTTP 80 rule, https:// instead of http://, or the wrong public IP (it changes after a stop and start).
2. You see Apache's test page or an error: the problem is inside. Check the service status, the folder and the file name.

Example 12: the test page means "no index.html yet".

[ec2-user@ip-172-31-xx-xx ~]$ curl -s http://localhost | grep title
<title>It works! Apache httpd</title>

Example 13: a page that isn't there.

[ec2-user@ip-172-31-xx-xx ~]$ curl http://localhost/about.html
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd">
<html><head>
<title>404 Not Found</title>
</head><body>
<h1>Not Found</h1>
<p>The requested URL was not found on this server.</p>
</body></html>

Example 14: just the headers. -I shows the status line and headers without the page:

[ec2-user@ip-172-31-xx-xx ~]$ curl -I http://localhost
HTTP/1.1 200 OK
Date: ...
Server: Apache/2.4.68 (Amazon Linux)
...
Content-Type: text/html; charset=UTF-8

200 OK means the page was found and sent.

Example 15: a common typo. Typing curl twice makes curl treat the second curl as a website name:

[ec2-user@ip-172-31-xx-xx ~]$ curl curl http://localhost
curl: (6) Could not resolve host: curl
<h1>Hello from my EC2 server</h1>

The first line is the failed "site" called curl; the second is your real page.

The usual mistakes, in order of how often they happen:

1. Port 80 not added in the Security Group.
2. Typing https:// instead of http://.
3. The file made in the home folder (~) instead of /var/www/html.
4. A wrong file name, like Index.html or index.htm. It must be exactly index.html.
5. Apache not started, or not started again after a reboot (use sudo systemctl enable httpd).

15. More pages, and a real page

Why. A site has more than one page, and developers give you much more than one line.

Example 16: a second page.

sudo nano about.html

Type, then save with Ctrl+X, Y, Enter:

<h1>About this site</h1>

Open it with the file name after the IP:

1. http://PUBLIC_IP/ still shows index.html.
2. http://PUBLIC_IP/about.html shows the new page.
3. Any name that isn't in the folder gives 404 Not Found.

Example 17: replace the home page with a full page. Open index.html again with sudo nano index.html, delete the old line and paste a complete page, for example:

<!DOCTYPE html>
<html>
<head>
  <title>My first website</title>
  <style>
    body { font-family: sans-serif; text-align: center; }
    h1 { color: #0b62a4; }
  </style>
</head>
<body>
  <h1>My first website</h1>
  <p>Hosted on Amazon EC2 with Apache.</p>
  <p>This page is served by Apache from /var/www/html.</p>
</body>
</html>

Save and refresh the browser. No restart is needed: Apache reads the file again for every request.

16. The whole setup on one page

What. Every command from this chapter, in order:

17. Try at home

Recap

Ravindra Bagale, trainer: linkedin.com/in/ravindra-bagale. The site name www.mysite.com, the pages and the user name ravi are examples for learning; PUBLIC_IP stands for your instance's public IPv4 address. The commands and messages were checked on a test copy of Amazon Linux 2023 (Apache httpd 2.4.68, nano 8.3). The hostname in the prompt, the dates and the version numbers are illustrative and change with updates.