Ssh Using Public Key

SSH stands for Secure Shell and is a cryptographic protocol based on the concept of public-private keys. We are using SSH with Git because it is much easier than typing your username and password. Public key authentication is a way of logging into an SSH/SFTPaccount using a cryptographic key rather than a password. If you use very strong SSH/SFTP passwords, your accounts are already safe from brute force attacks. However, using public key authentication provides many benefits when working with multiple developers. For example, with SSH keys you can 1. Allow multiple developers to log i.

-->

This article is for Windows users who want to create and use secure shell (SSH) keys to connect to Linux virtual machines (VMs) in Azure. You can also generate and store SSH keys in the Azure portal to use when creating VMs in the portal.

Public Key Private Key Ssh

To use SSH keys from a Linux or macOS client, see the quick steps. For a more detailed overview of SSH, see Detailed steps: Create and manage SSH keys for authentication to a Linux VM in Azure.

Overview of SSH and keys

SSH is an encrypted connection protocol that allows secure sign-ins over unsecured connections. SSH is the default connection protocol for Linux VMs hosted in Azure. Although SSH itself provides an encrypted connection, using passwords with SSH still leaves the VM vulnerable to brute-force attacks. We recommend connecting to a VM over SSH using a public-private key pair, also known as SSH keys.

The public-private key pair is like the lock on your front door. The lock is exposed to the public, anyone with the right key can open the door. The key is private, and only given to people you trust because it can be used to unlock the door.

The public key is placed on your Linux VM when you create the VM.
The private key remains on your local system. Protect this private key. Do not share it.

When you connect to your Linux VM, the VM tests the SSH client to make sure it has the correct private key. If the client has the private key, it's granted access to the VM.

Depending on your organization's security policies, you can reuse a single key pair to access multiple Azure VMs and services. You do not need a separate pair of keys for each VM.

Your public key can be shared with anyone, but only you (or your local security infrastructure) should have access to your private key.

Supported SSH key formats

Azure currently supports SSH protocol 2 (SSH-2) RSA public-private key pairs with a minimum length of 2048 bits. Other key formats such as ED25519 and ECDSA are not supported.

SSH clients

Recent versions of Windows 10 include OpenSSH client commands to create and use SSH keys and make SSH connections from PowerShell or a command prompt. This is the easiest way to create an SSH connection to your Linux VM, from a Windows computer.

You can also use Bash in the Azure Cloud Shell to connect to your VM. You can use Cloud Shell in a web browser, from the Azure portal, or as a terminal in Visual Studio Code using the Azure Account extension.

You can also install the Windows Subsystem for Linux to connect to your VM over SSH and use other native Linux tools within a Bash shell.

Create an SSH key pair

Create an SSH key pair using the ssh-keygen command. Enter a filename, or use the default shown in parenthesis (for example C:Usersusername/.ssh/id_rsa). Enter a passphrase for the file, or leave the passphrase blank if you do not want to use a passphrase.

Create a VM using your key

To create a Linux VM that uses SSH keys for authentication, provide your SSH public key when creating the VM.

Using the Azure CLI, you specify the path and filename for the public key using az vm create and the --ssh-key-value parameter.

With PowerShell, use New-AzVM and add the SSH key to the VM configuration using`. For an example, see Quickstart: Create a Linux virtual machine in Azure with PowerShell.

If you do a lot of deployments using the portal, you might want to upload your public key to Azure, where it can be easily selected when creating a VM from the portal. For more information, see Upload an SSH key.

Connect to your VM

With the public key deployed on your Azure VM, and the private key on your local system, SSH to your VM using the IP address or DNS name of your VM. Replace azureuser and 10.111.12.123 in the following command with the administrator user name, the IP address (or fully qualified domain name), and the path to your private key:

If you configured a passphrase when you created your key pair, enter the passphrase when prompted.

If the VM is using the just-in-time access policy, you need to request access before you can connect to the VM. For more information about the just-in-time policy, see Manage virtual machine access using the just in time policy.

Next steps

For information about SSH keys in the Azure portal, see Generate and store SSH keys in the Azure portal to use when creating VMs in the portal.
For detailed steps, options, and advanced examples of working with SSH keys, see Detailed steps to create SSH key pairs.
You can also use PowerShell in Azure Cloud Shell to generate SSH keys and make SSH connections to Linux VMs. See the PowerShell quickstart.
If you have difficulty using SSH to connect to your Linux VMs, see Troubleshoot SSH connections to an Azure Linux VM.

Did you know you can passwordless SSH? Here's how, and how to decide whether you should.

If you interact regularly with SSH commands and remote hosts, you may find that using a key pair instead of passwords can be convenient. Instead of the remote system prompting for a password with each connection, authentication can be automatically negotiated using a public and private key pair.

The private key remains secure on your own workstation, and the public key gets placed in a specific location on each remote system that you access. Your private key may be secured locally with a passphrase. A local caching program such as ssh-agent or gnome-keyring allows you to enter that passphrase periodically, instead of each time you use the key to access a remote system.

[ Free download: Advanced Linux commands cheat sheet. ]

Generating a key pair and propagating the public key

Generating your key pair and propagating your public key is simpler than it sounds. Let’s walk through it.

Generating the key

The minimum effort to generate a key pair involves running the ssh-keygen command, and choosing the defaults at all the prompts:

The default location to store the keys is in the ~/.ssh directory, which will be created if it does not exist:

Allowing this command to create the directory also ensures that the owner and permissions are set correctly. Some applications will not use keys if the permissions to the private key are too open.

The file ending in .pub is the public key that needs to be transferred to the remote systems. It is a file containing a single line: The protocol, the key, and an email used as an identifier. Options for the ssh-keygen command allow you to specify a different identifier:

After generating the key pair, the ssh-keygen command also displays the fingerprint and randomart image that are unique to this key. This information can be shared with other people who may need to verify your public key.

Later you can view these with:

The -l option lists the fingerprint, and the -v option adds the ASCII art.

Propagating the public key to a remote system

If password authentication is currently enabled, then the easiest way to transfer the public key to the remote host is with the ssh-copy-id command. If you used the default name for the key all you need to specify is the remote user and host:

Following the instructions from the output, verify that you can connect using the key pair. If you implemented a passphrase, you will be prompted for the passphrase to use the private key:

Examine the resulting authorized key file. This is where the public key was appended. If the directory or file did not exist, then it was (or they were) created with the correct ownership and permissions. Each line is a single authorized public key:

To revoke access for this key pair, remove the line for the public key.

There are many other options that can be added to this line in the authorized key file to control access. These options are usually used by administrators placing the public keys on a system with restrictions. These restrictions may include where the connection may originate, what command(s) may be run, and even a date indicating when to stop accepting this key. These and more options are listed in the sshd man page.

Changing the passphrase

If you need to change a passphrase on your private key or if you initially set an empty passphrase and want that protection at a later time, use the ssh-keygen command with the -p option:

You can add additional options to specify the key (-f), and the old (-P) or new (-N) passphrases on the command line. Remember that any passwords specified on the command line will be saved in your shell history.

See the ssh-keygen man page for additional options.

Rotating keys

While the public key by itself is meant to be shared, keep in mind that if someone obtains your private key, they can then use that to access all systems that have the public key. These key pairs also do not have a period of validity like GNU Privacy Guard (GPG) keys or public key infrastructure (PKI) certificates.

If you have any reason to suspect that a private key has been stolen or otherwise compromised, you should replace that key pair. The old public key has to be removed from all systems, a new key has to be generated with ssh-keygen, and the new public key has to be transferred to the desired remote systems.

If you are rotating keys as a precaution and without any concern of compromise, you can use the old key pair to authenticate the transfer of the new public key before removing the old key.

Is using empty passphrases ever a good idea?

There are several things to think about when considering an empty passphrase for your SSH private key.

How secure is the private key file?

If you tend to work from multiple client systems and want to either have multiple copies of your key or keep a copy on removable media, then it really is a good idea to have a passphrase on the private key. This practice is in addition to protecting access to the key file with encrypted media.

However, if you have only one copy of the private key and it is kept on a system that is well secured and not shared, then having a passphrase is simply one more level of protection just in case.

Remember that changing the passphrase on one copy does not change the passphrase on other copies. The passphrase is simply locking access to a specific key file.

Why do think you need an empty passphrase?

There are cases for keys with empty passphrases. Some utilities that need to automatically transfer files between systems need a passwordless method to authenticate. The kdump utility, when configured to dump the kernel to a remote system using SSH, is one example.

Another common use is to generate a key pair for a script that is designed to run unattended, such as from a cron job.

How about a middle ground alternative?

By itself, a passphrase-protected private key requires the passphrase to be entered each time the key is used. This setup does not feel like passwordless SSH. However, there are caching mechanisms that allow you to enter the key passphrase once and then use the key over and over without reentering that passphrase.

More Linux resources

OpenSSH comes with an ssh-agent daemon and an ssh-add utility to cache the unlocked private key. The GNOME desktop also has a keyring daemon that stores passwords and secrets but also implements an SSH agent.

The lifetime of the cached key can be configured with each of the agents or when the key is added. In many cases, it defaults to an unlimited lifetime, but the cache is cleared when the user logs out of the system. You will be prompted for the passphrase only once per login session.

If there is a scheduled application that needs to run outside of a user login session, it may be possible to use a secret or other password manager to automate the unlocking of the key. For example, Ansible Tower stores credentials in a secure database. This database includes an SSH private key used to connect to the remote systems (managed nodes), and any passphrases necessary for those private keys. Once those credentials are stored, a job can be scheduled to run a playbook on a regular schedule.

Automating propagation

A centralized identity manager such as FreeIPA can assist with key propagation. Upload the public key to the server as an attribute of a user account, and then propagate it to the hosts in the domain as needed. FreeIPA can also provide additional host-based access control for where a key may be used.

Keys can also be distributed using Ansible modules. The openssh_keypair module uses ssh-keygen to generate keys and the authorized_key module adds and removes SSH authorized keys for particular user accounts.

Wrapping up

SSH key pairs are only one way to automate authentication without passwords. Using the Generic Security Services Application Program Interface (GSSAPI) authentication is also common when trying to reduce the use of passwords on a network with centralized user management. SSH key pairs are the easier option to implement when single sign-on (SSO) is not already available.