How to secure a VPS
Objective
When you order your VPS, you can choose a distribution or operating system to install. The server is therefore ready to use after delivery but it will be up to you as the administrator to implement measures to ensure the security and stability of your system.
This guide provides some general tips for securing a GNU/Linux-based server.
OVHcloud is providing you with services for which you are responsible, with regard to their configuration and security. Since we have no administrative access to your devices, it is your responsibility to manage the software and to ensure they function correctly.
This guide is designed to help you with the most common tasks. Nevertheless, we recommend that you contact a specialist service provider if you have difficulties or doubts concerning the administration, usage or implementation of security measures on a server.
Requirements
- A Virtual Private Server in your OVHcloud account
- Administrative access (sudo) via SSH to your server
Instructions
Bear in mind that this is a general guide based on Ubuntu, Debian and CentOS operating systems. Some commands need to be adapted to the distribution or operating system you are using and some tips will advise you to use third-party tools. Please refer to the official documentation for these applications if you require assistance.
If you are configuring your first OVHcloud VPS, we recommend to consult our guide on getting started with a VPS before continuing.
The following examples presume that you are logged in as a user with elevated permissions.
Table of Contents:
- Updating your system
- Create and use an SSH key
- Changing the default SSH listening port
- Creating a user with restricted rights
- Configuring the internal firewall (iptables)
- Installing Fail2ban
- Configuring the OVHcloud Network Firewall
- Backing up your system and your data
Updating your system
Developers of distributions and operating systems offer frequent software package updates, very often for security reasons. Ensuring that your distribution or operating system is updated is a key point for securing your VPS.
This update will take place in two steps:
- Updating the package list:
- Updating the actual packages:
The command is identical to Ubuntu because Debian and Ubuntu both use apt.
On CentOS, the command to update the operating system uses yum or dnf, depending on the version.
This operation needs to be performed regularly to keep a system up-to-date.
Create and use an SSH key
SSH key authentication is one of the most effective methods for securing access to your VPS.
Unlike password authentication, it relies on a pair of cryptographic keys and significantly reduces the risk of brute-force attacks.
We strongly recommend setting up an SSH key during your first connection to your server, and then prioritising this method for your administrative access.
Depending on your environment and the tool you use to connect to your VPS, refer to one of the following guides:
- How to create and use authentication keys for SSH connections to OVHcloud servers
- Tutorial - How to use PuTTY for SSH connections and authentication
These guides detail the steps to:
- generate an SSH key pair;
- deploy the public key on your server;
- securely connect via SSH.
Once SSH key authentication is configured and working, you can go further by enhancing the SSH service configuration, for example by changing the listening port or disabling password authentication.
Changing the default SSH listening port
Before making any changes to the SSH service, ensure you have a working SSH key access to avoid losing access to your server.
For this section, the following command lines are the same for Ubuntu, Debian, and CentOS.
One of the first things to do on your server is configuring the SSH service's listening port. It is set to port 22 by default, therefore server hacking attempts by robots will target this port. Modifying this setting by using a different port is a simple measure to harden your server against automated attacks.
To do this, modify the service configuration file with a text editor of your choice (nano used in this example):
Find the following or similar lines:
Replace the number 22 with the port number of your choice. Please do not enter a port number already used on your system. To be safe, use a number between 49152 and 65535.
Alternatively, you can view the ports assigned on your system with the following command:
If the line is "commented out" (i.e. if it is preceded by a "#") as shown in the example above, make sure to remove the "#" before saving the file so that the change takes effect. Example:
If a firewall is configured on your operating system (UFW or iptables), you must adjust its settings to allow traffic on the new port before restarting the service. If you are using iptables, refer to this guide: Configuring the firewall on Linux with iptables. If no firewall is configured by default, restart the service.
Save and exit the configuration file.
Restart the service:
This should be sufficient to apply the changes. Alternatively, reboot the VPS (sudo reboot).
For Ubuntu 24.04 and later
For the latest Ubuntu versions, the SSH configuration is now managed in the ssh.socket file.
To update the SSH port, edit the Listenstream line in the configuration file with a text editor of your choice (nano used in this example):
Your file should resemble the following examples, depending on the version of Ubuntu you have installed:
or
Save your changes and run the following commands:
Restart the service:
Remember that you will have to indicate the new port any time you establish an SSH connection to your server:
Example:
If you are locked out of your system, you can use our rescue mode environment to revert your changes.
Creating a user with restricted rights
In general, tasks that do not require root privileges should be performed via a standard user. Please refer to the information in this guide for details.
Configuring the internal firewall (iptables)
Common GNU/Linux distributions come with a firewall service named iptables. By default, this service does not have any active rules. You can verify this by typing the following command:
You can learn more about iptables in our Firewall guide.
It is recommended that you create and adjust firewall rules according to your needs. For more detailed information on the variety of manipulations that are possible, please refer to the relevant section in the official documentation of the distribution used.
Installing Fail2ban
Fail2ban is an intrusion prevention software framework designed to block IP addresses from which bots or attackers try to penetrate your system. This software package is recommended, even essential in some cases, to guard your server against "Brute Force" or "Denial of Service" attacks.
To install the software package, use the following command:
On CentOS 7 and CentOS 8 (or RHEL), first install the EPEL repository (Extra Packages for Enterprise Linux), then Fail2ban:
You can customize the Fail2ban configuration files to protect services that are exposed to the public Internet from repeated login attempts.
As recommended by Fail2ban, create a local configuration file for your services by copying the "jail" file:
Then open the file with a text editor:
Be certain to read the information at the top of the file, especially the comments under [DEFAULT].
The [DEFAULT] settings are global and will therefore be applied to all services that are set to enabled in this file.
It is important to know that the global settings will be taken into account only if there are no differing values set in the services sections (JAILS) further below in the file.
For example, consider these lines under [DEFAULT]:
This means that an IP address from which a host tries to connect will be blocked for ten minutes after the fifth unsuccessful login attempt.
However, all settings specified by [DEFAULT] and in subsequent sections stay disabled unless the line enabled = true is added for a service (listed below # JAILS).
As an example of usage, having the following lines in the section [sshd] will activate restrictions only for the OpenSSH service:
In this example, any SSH login attempt that fails three times within five minutes will result in an IP ban period of 30 minutes.
You can replace "ssh" with the actual port number in case you have changed it.
The best practice approach is to enable Fail2ban only for the services that are actually running on the server. Each customized setting added under # JAILS will then be prioritized over the defaults.
Once you have completed your changes, save the file and close the editor.
Restart the service to make sure it runs with the customization applied:
1. Recommended command with systemctl:
2. Command with service (legacy method, still compatible):
Fail2ban has many settings and filters for customization as well as preset options, for example when you want to add a layer of protection to an Nginx web server.
For any additional information and recommendations concerning Fail2ban, please refer to the official documentation of this tool.
Configuring the OVHcloud Network Firewall
OVHcloud solutions include the option of enabling a firewall at the entry point to the infrastructure, called the Network Firewall. Configuring it correctly allows connections to be blocked before they even arrive on your server.
Please refer to the Network Firewall guide if you would like to activate it.
Backing up your system and your data
The concept of security is not limited to protecting a system against attacks.
Securing your data is a key element, which is why OVHcloud offers you several backup options as a service:
- The
Snapshotoption allows you to create a manual snapshot. - The
Automated Backupoption enables you to keep regular backups of your VPS (excluding additional disks).
You can find all information on the available backup solutions for your service on the product page and in the respective guides.
Go further
How to create and use SSH keys
Configuring the firewall on Windows
Configuring the firewall on Linux with iptables
Join our community of users.
-
Secure Shell (SSH) : un protocole de réseau sécurisé utilisé pour établir des connexions entre un client et un serveur. Il permet d'exécuter des commandes à distance de manière sécurisée. ↩