Ssh-keygen Save Location

[Roselee Kruppa]

SSH introduced public key authentication as a more secure alternative to the older .rhosts authentication. It improved security by avoiding the need to have password stored in files, and eliminated the possibility of a compromised server stealing the user's password.

However, SSH keys are authentication credentials just like passwords. Thus, they must be managed somewhat analogously to user names and passwords. They should have a proper termination process so that keys are removed when no longer needed.

First, the tool asked where to save the file. SSH keys for user authentication are usually stored in the user's .ssh directory under the home directory. However, in enterprise environments, the location is often different. The default key file name depends on the algorithm, in this case id_rsa when using the default RSA algorithm. It could also be, for example, id_dsa or id_ecdsa.

Then it asks to enter a passphrase. The passphrase is used for encrypting the key, so that it cannot be used even if someone obtains the private key file. The passphrase should be cryptographically strong. Our online random password generator is one possible tool for generating strong passphrases.

rsa - an old algorithm based on the difficulty of factoring large numbers. A key size of at least 2048 bits is recommended for RSA; 4096 bits is better. RSA is getting old and significant advances are being made in factoring. Choosing a different algorithm may be advisable. It is quite possible the RSA algorithm will become practically breakable in the foreseeable future. All SSH clients support this algorithm.

dsa - an old US government Digital Signature Algorithm. It is based on the difficulty of computing discrete logarithms. A key size of 1024 would normally be used with it. DSA in its original form is no longer recommended.

ecdsa - a new Digital Signature Algorithm standarized by the US government, using elliptic curves. This is probably a good algorithm for current applications. Only three key sizes are supported: 256, 384, and 521 (sic!) bits. We would recommend always using it with 521 bits, since the keys are still small and probably more secure than the smaller keys (even though they should be safe as well). Most SSH clients now support this algorithm.

Once the public key has been configured on the server, the server will allow any connecting user that has the private key to log in. During the login process, the client proves possession of the private key by digitally signing the key exchange.

ssh-agent is a program that can hold a user's private key, so that the private key passphrase only needs to be supplied once. A connection to the agent can also be forwarded when logging into a server, allowing SSH commands on the server to use the agent running on the user's desktop.

The host keys are usually automatically generated when an SSH server is installed. They can be regenerated at any time. However, if host keys are changed, clients may warn about changed keys. Changed keys are also reported when someone tries to perform a man-in-the-middle attack. Thus it is not advisable to train your users to blindly accept them. Changing the keys is thus either best done using an SSH key management tool that also changes them on clients, or using certificates.

OpenSSH does not support X.509 certificates. Tectia SSH does support them. X.509 certificates are widely used in larger organizations for making it easy to change host keys on a period basis while avoiding unnecessary warnings from clients. They also allow using strict host key checking, which means that the clients will outright refuse a connection if the host key has changed.

OpenSSH has its own proprietary certificate format, which can be used for signing host certificates or user certificates. For user authentication, the lack of highly secure certificate authorities combined with the inability to audit who can access a server by inspecting the server makes us recommend against using OpenSSH certificates for user authentication.

However, OpenSSH certificates can be very useful for server authentication and can achieve similar benefits as the standard X.509 certificates. However, they need their own infrastructure for certificate issuance.

It is easy to create and configure new SSH keys. In the default configuration, OpenSSH allows any user to configure new keys. The keys are permanent access credentials that remain valid even after the user's account has been deleted.

In organizations with more than a few dozen users, SSH keys easily accumulate on servers and service accounts over the years. We have seen enterprises with several million keys granting access to their production servers. It only takes one leaked, stolen, or misconfigured key to gain access.

In any larger organization, use of SSH key management solutions is almost necessary. SSH keys should also be moved to root-owned locations with proper provisioning and termination processes. For more information, see how to manage SSH keys. A widely used SSH key management tool for OpenSSH is Universal SSH Key Manager.

Practically all cybersecurity regulatory frameworks require managing who can access what. SSH keys grant access, and fall under this requirement. This, organizations under compliance mandates are required to implement proper management processes for the keys. NIST IR 7966 is a good starting point.

It is important to ensure there is enough unpredictable entropy in the system when SSH keys are generated. There have been incidents when thousands of devices on the Internet have shared the same host key when they were improperly configured to generate the key without proper randomness.

On general purpose computers, randomness for SSH key generation is usually not a problem. It may be something of an issue when initially installing the SSH server and generating host keys, and only people building new Linux distributions or SSH installation packages generally need to worry about it.

Our recommendation is to collect randomness during the whole installation of the operating system, save that randomness in a random seed file. Then boot the system, collect some more randomness during the boot, mix in the saved randomness from the seed file, and only then generate the host keys. This maximizes the use of the available randomness. And make sure the random seed file is periodically updated, in particular make sure that it is updated after generating the SSH host keys.

Many modern general-purpose CPUs also have hardware random number generators. This helps a lot with this problem. The best practice is to collect some entropy in other ways, still keep it in a random seed file, and mix in some entropy from the hardware random number generator. This way, even if one of them is compromised somehow, the other source of randomness should keep the keys secure.

Available entropy can be a real problem on small IoT devices that don't have much other activity on the system. They may just not have the mechanical randomness from disk drive mechanical movement timings, user-caused interrupts, or network traffic. Furthermore, embedded devices often run on low-end processors that may not have a hardware random number generator.

We at SSH secure communications between systems, automated applications, and people. We strive to build future-proof and safe communications for businesses and organizations to grow safely in the digital world.

My Git setup runs fine on Linux, but when I try to set things up under Windows (using Git for Windows and TortoiseGit), I don't know where to put my private SSH key (or, better still, how to tell ssh where it's located). I'm using the standard ssh.exe option during installation of Git for Windows. The setup runs fine if I allow password authentication (in lieu of RSA) on the server.

If you are running msysgit (I am assuming you are) and are looking to run Git Bash (I recommend it over TortoiseGit, but I lean to the CLI more than GUI now), you need to figure out what your home directory is for Git Bash by starting it then type pwd (On Windows 7, it will be something like C:\Users\phsr I think). While you're in Git Bash, you should mkdir .ssh.

After you have the home directory, and a .ssh folder under that, you want to open PuTTYgen and open the key (.ppk file) you have previously created. Once your key is open, you want to select Conversions -> Export OpenSSH key and save it to HOME\.ssh\id_rsa. After you have the key at that location, Git Bash will recognize the key and use it.

It made my day and fixed the issue with Git provided that your private key is not password protected. If you want to use ssh-agent, then you can probably run ssh-agent cmd.exe (although I've never done that) and the ssh-add as usual.

If this does not work correctly, using plink can probably be achieved by tweaking GIT_SSH. Refer to all the SVN + ssh tutorials; this is basically the same plumbing you need to setup.

Here's an example of a Git configuration for Bitbucket that will use the correct username, and a key other than the default key (in case you maintain one key for SSH connections, and another for Git accounts).

Plus if you plan using ssh-keyscan host.com > known_hosts from OpenSSH, be careful because piping output from stdout of keyscan (on Windows) changes encoding to UCS-2, but OpenSSH can read only UTF-8! So make sure to change the known_hosts file encoding.

Your private key needs to be added to the SSH agent on your workstation.How you achieve this may depend on what git client you are using, however puTTY and its associated agent (pageant) might do the trick for you, here's the link to the official binaries and source:

The most efficient way is using Pageant because it will let you write the passphrase only once at the beginning of the session instead of every push. All answers here were too short, so I'll post a detailed guide here:

When mixing GitHub for Windows and Git GUI for Windows, you might run into issues where Git GUI keeps prompting you for a username and password. The cure for this is to change the remote URL from https: (which is what GitHub for Windows creates) to the Git protocol. In the .git directory in the configuration file, find:

c80f0f1006