A few days ago I bought a very cheap Virtual Private Server (VPS) – check my post here: https://www.ajni.it/2020/03/quick-tip-cheap-private-servers-on-the-cloud/

It was very cheap (4$ or 3.75€ annually), but with a lot of gotchas.

One of them is Ubuntu 18.04 Minimal, which means a lot of packages will not be pre-installed, causing a lot of pain when installing services like in my example OpenVPN.

Here is how I managed to install OpenVPN on Ubuntu 18.04 Minimal.

Updating the system:

apt update

apt upgrade

Install OpenVPN

wget https://git.io/vpn -O openvpn-install.sh && bash openvpn-install.sh

The first problem occurs with the root CA certificates:

Install the root certificates in order to trust them:

apt-get install ca-certificates

After re-running the command, another error shows up:

Install the next package (iptables):

apt-get install iptables

And finally, the OpenVPN setup can be run:

I had to set a custom port, because only specific ones were NAT’d to my server. You might leave the port to default. I am also using 1.1.1.1 for DNS.

After the setup is finished, a configuration file will be created. This file contains the public certificates and private key that are mandatory for the connection. It can be imported into the OpenVPN client (Windows) through the GUI.

On Linux, a simple

openvpn configfile.ovpn

does the trick.

By default, Linux systems allow both password-based and key-based authentication over SSH. If you have a server with SSH open to the world, password-based authentication shouldn’t be allowed at all.

To disable password-based authentication, edit the SSH config file:

nano /etc/ssh/sshd_config

Add the following lines:

PasswordAuthentication no

PubkeyAuthentication yes

Now generate a new private/public key pair:

ssh-keygen

id_rsa is your private key

id_rsa.pub is the public certificate thumbprint that must be added to ~/.ssh/authorized_keys

nano ~/.ssh/authorized_keys

The SSH service must be restarted.

service ssh restart

Now you can connect to your server with key-based authentication only. If connecting from a Linux system the file’s permissions must be set to 600.

chmod 600 id_rsa

ssh -i id_rsa ip@username

If you like using Putty, you’ll have to load the file with PuttyGen and save the private key as .ppk.

PuttyGen can be downloaded here: https://www.chiark.greenend.org.uk/~sgtatham/putty/latest.html

Links:

https://askubuntu.com/questions/346857/how-do-i-force-ssh-to-only-allow-users-with-a-key-to-log-in

Today I stumbled upon this very useful site that helps you find very cheap virtual private servers (VPS) around the globe. The public IPv4 address is shared and NAT'd across multiple servers. If that's no problem for you, the cheapest servers are 0.15€ a month with 128MB RAM and 1 vCPU.

Also, some public IP addresses are blocked in China and/or Russia.

Have fun!

https://www.serverhunter.com/

I got tasked with installing a brand new Windows Server 2019 with the Remote Desktop Services (RDS) role. It was a pretty straightforward installation, but minor things might work differently compared to previous versions of Windows Server (I was migrating off Windows Server 2012 R2).

Start by adding the RDS role through Server Manager.

Select the RDS installation:

I went for the Quick Start because my deployment is fairly basic.

Session-based deployment enables multi-session support on the server.

The server should be automatically selected.

The three roles (RD Connection Broker, RD Web Access and RD Session Host) will be installed.

After the installation a license warning will be shown in the Notification Center.

Two things are needed in order for licensing to work properly: The license server and licensing mode.
In a production environment, usually there is a separate server hosting the RDS Licensing service.

Server Manager > Remote Desktop Services > Overview > Tasks > Edit Deployment properties

I had problems with the licensing mode not being applied properly. This registry key worked wonderfully though:

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Terminal Server\RCM\Licensing Core
DWORD LicensingMode
4 = Per user
2 = Per Device

The license server can be also set through the registry:

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\TermService\Parameters\LicenseServers
REG_MULTI_SZ SpecifiedLicenseServers = license.ajni.lab

With RD Licensing Diagnoser you can check for potential errors (can be opened through Server Manager > Tools > Remote Desktop Services).

Now on to the last step: Create a custom device collection.

Server Manager > Remote Desktop Services > Collections > Tasks > Create Session Collection

This is also pretty straight forward stuff.

Make sure you select the server by moving it to the right with the arrow.

In a production environment a custom group should be used to control the number of permitted users.

User profile disks were not needed in my environment.

After the creation, there are some things that should be changed in the Collection properties (Server Manager > Remote Desktop Services > Collections > Collection Name):

These are my specific settings, you should change the parameters based on your experience or leave them at their default values.

Older clients might have problems with these security settings (like Network Level Authentication - NLA)

Do not forget to apply the changes.

Bonus:

If you have specific AD user attributes, like the home folder or program auto-start, they will not work because of changes made to RDS 2016/2019. You can read this article from Microsoft's website: https://support.microsoft.com/en-us/help/3200967/changes-to-remote-connection-manager-in-windows-server

Following registry entries will tell the Remote Desktop Session Host (RDSH) to query AD DS for RDP profile settings:

HKEY_LOCAL_MACHINE\SOFTWARE\Policies\Microsoft\Windows NT\Terminal Services
Name: fQueryUserConfigFromDC
Type: Reg_DWORD
Value: 1

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Terminal Server\WinStations\RDP-tcp
Name: fQueryUserConfigFromDC
Type: Reg_DWORD

A quick post on how to change the Windows display language with Powershell. You might use these commands based on any logic that determines the user's location/language. For instance, I created a script that gets executed on logon and sets the language based on some criteria (maybe an Active-Directory group or attribute).

Set-Culture en-US
Set-WinSystemLocale -SystemLocale en-US
Set-WinUILanguageOverride -Language en-US
Set-WinUserLanguageList en-US -Force
Set-WinHomeLocation -GeoId 244

You can find the right GeoID on Microsoft's website