SSH No Ports

Installation Guide

Learn more about installation requirements

Read the Installation FAQ

Prerequisites

Please note: We strongly discourage installing SSH No Ports as root. If you don't have another user account available, we recommend creating one first. (A quick web search will help you find steps for your operating system.)

Before you begin, you'll need to set up curl and sshd on both your client and your remote device.

curl
Simply type the following command into your terminal:

curl

You should be greeted with this message: "curl: try 'curl --help' for more information"

If you see something like "command not found" a quick web search will show you how to install curl on your machine.

Have both your client and device ready—prior to installation make sure the following is installed and configured accordingly:

sshd
SSH No Ports needs to be able to SSH to localhost. To test for this, start by running the following command:

ssh 0

If it asks for a password, or greets you with an SSH session, you are good to go.

If you see something like: "ssh: connect to host 0.0.0.0 port 22: Connection refused" then you will either need to install or enable sshd. The method can vary depending on the platform; some common solutions to this are:

macOS:
Turn on remote login in System Settings

Linux Desktop

sudo apt install openssh-server
sudo service ssh restart


Systemd based servers

sudo systemctl enable sshd.service

The following steps require that you alternate between client and device. It is best if you have two terminals open, one for each.

1. Get your addresses

Once you have the above prerequisites setup on both the client and device, get your addresses by purchasing SSH No Ports from my.noports.com. You will receive both a client and device address.

All directions in Section 2 refer to the device.

2.1 Set up the device

Run the following command to install the device daemon:

bash -c "$(curl -fsSL https://getsshnpd.noports.com)" -- --args "-s -u -v"

The program will ask you for some input. First enter your client and device addresses from step 1. You will also need to pick a unique device name using only alphanumeric characters and “_” with a maximum length of 15 characters.

Client address (e.g. @alice_client): <client address>
Device address (e.g. @alice_device): <device address>
Device name: <device name>


Security notice: passing `-- –args “-s -u -v”` will append flags to the sshnpd binary call where `-s` will add any new ssh public keys to the device’s authorized_keys, `-u` will share the sensitive device information (such as the username) with the client address, and `-v` will enable logging.

Example:

Client address (e.g. @alice_client): @alice_client
Device address (e.g. @alice_device): @alice_device
Device name: raspberry_pi

OR

If you would like to automate this in a script, pass the input as arguments like so:

bash -c "$(curl -fsSL https://getsshnpd.noports.com)" -- -c <client address> -d <device address> -n <device name>


Example:

bash -c "$(curl -fsSL https://getsshnpd.noports.com)" -- -c @alice_client -d @alice_device -n raspberry_pi

2.2 Add the binaries to the PATH

This step is optional, but highly recommended. If you skip this step, you must prefix all of the SSH No Ports binaries (at_activate, sshnp, sshnpd, etc.) with the following path: “~/.local/bin/”.

Add the following line to your shell’s rc file to add all of the newly installed binaries to your path:

export PATH="$HOME/.local/bin:$PATH";

If you use bash (most Linux distributions by default) add the above line to '~/.bashrc'.
If you use zsh (most macOS distributions by default) add the above line to '~/.zshrc'.

All directions in Sections 3 refer to the client.

3.1 Set up the client

Run the following command to install the client:

bash -c "$(curl -fsSL https://getsshnp.noports.com)"

The program will ask you for some input. First enter your client and device address from Step 1. You will also need to pick a default region, ideally this is the closest to your physical location.

Client address (e.g. @alice_client): <client address>
Device address (e.g. @alice_device): <device address>
Pick your default region:  
am: Americas
ap: Asia Pacific
eu: Europe
> <region code>


Example:

Client address (e.g. @alice_client): @alice_client
Device address (e.g. @alice_device): @alice_device
Pick your default region:  
am: Americas  
ap: Asia Pacific  
eu: Europe
> am

OR

If you would like to automate this in a script, pass the input as arguments like so:

bash -c "$(curl -fsSL https://getsshnp.noports.com)" -- -c <client address> -d <device address> -h <region code>


Example:

bash -c "$(curl -fsSL https://getsshnp.noports.com)" -- -c @alice_client -d @alice_device -h am

3.2 Add the binaries to the PATH


This step is optional, but
highly recommended. If you skip this step, you must prefix all of the SSH No Ports binaries (at_activate, sshnp, sshnpd, etc.) with the following path: “~/.local/bin/”.

Add the following line to your shell’s rc file to add all of the newly installed binaries to your path:

export PATH="$HOME/.local/bin:$PATH";

If you use bash (most Linux distributions by default) add this line to '~/.bashrc'.
If you use zsh (most macOS distributions by default) add this line to '~/.zshrc'.

3.3 Activate the client address

Before the sshnp client  can be used, it must be activated. Use the at_activate utility that comes with sshnp. It requires the device address as input (you must include the “@” prefix).

at_activate -a <client address>

Example:

at_activate -a @alice_client

Then a one-time password will be sent to your email address to confirm activation. Note that this part cannot be automated for security reasons.

3.4 Activate the device address

Exactly like the last step, activate the device address on your client machine. Your client machine will hold both the client and device keys. It requires the device address as input (you must include the "@" prefix). Successful activation will result in your device's keys being generated in ~/.atsign/keys.

at_activate -a <device address>

Example:

at_activate -a @alice_device

Then a one-time password will be sent to your email address to confirm activation. Note that this part cannot be automated for security reasons.

3.5 Back up the device keys

[!] This is a temporary step.

For features that are currently being worked on, this is a temporary step. The device keys should from now on be activated on your client machine, and future Atsign features will allow you to authenticate other devices from your client. For now, secure copy your device keys from your client machine to your device machine.

On your client machine, SCP (Secure Copy) the device's keys (which can be found on the client machine's ~/.atsign/keys/directory) into the same directory on your device's machine (~.atsign/keys). It requires the device address and the user and host of your device as input.

scp ~/.atsign/keys/<device address>_key.atKeys <user>@<host>:<device home directory>/.atsign/keys/<device address>_key.atKeys

Example:

scp ~/.atsign/keys/@alice_device_key.atKeys
alice@123.456.789:/alice/.atsign/keys/@alice_device_key.atKeys


It is important that the keys are secure copied to the path: <home directory>/.atsign/keys on the device machine. The ~/.atsign/keys/ directory should have already been created on the device's machine after running the first install command.

All directions in Section 4 refer to the device.

4. Connect to your device

[!] This is a temporary step.

Versions <=v3.3.0 will experience a bug which requires a manual restart sshnpd on the device. This bug only affects the initial run of sshnpd, and will be fixed in the next release.

killall -u "$(whoami)" -r "sshnpd$"

All directions in Section 5 refer to the client.

5. Connect to your device from your client

$(sshnp@<device address> -d <device name>)

Example:

$(sshnp@alice_device -d raspberry_pi)

Updating your installation

Device

To update the device, run the following command:

bash -c "$(curl -fsSL https://getsshnpd.noports.com)" -- --update

This will update all services installed under the current user.

Client


To update the client, run the following command:

bash -c "$(curl -fsSL https://getsshnp.noports.com)" -- --update

This will update the sshnp client for the current user.

Installation FAQ

1. How do I resolve an SSHAuthAbortError (connection closed before authentication)?

Generally, this error is caused because sshd is not running on one of the machines. SSH No Ports requires sshd to be enabled for localhost connections on both devices. To test whether the device is running sshd on localhost, run the following command on the device:

ssh 0

If you receive this error, "ssh: connect to host 0.0.0.0 port 22: Connection refused" then sshd is disabled and you can follow the guide below to enable it.

If you are able to successfully run "ssh 0" on both devices and still experience issues, please contact us for further assistance.

Then a one-time password will be sent to your email address to confirm activation. Note that this part cannot be automated for security reasons.

How to enable sshd

macOS: Simply enable Remote Login in System Settings.

Linux: Solutions can vary. A quick web search will help you find the appropriate solution for your specific distribution.

Recommended sshd configuration

In addition, we recommend adding the following configuration to the /ect/ssh/sshd_config file:

ListenAddress 127.0.0.1
PasswordAuthentication no

Once you've added this configuration, restart sshd.