Setting the SSH host keys offered to the clients
By default, SPS automatically generates the private host keys for all supported host key algorithms when creating a new connection. You only have to make the settings described below if you would like to use your own host keys, or you would like to remove certain host keys.
The following describes how to upload or paste the private part of the SSH host key. SPS will offer the host keys to the clients.
To set the SSH host keys offered to the clients
-
Navigate to SSH Control > Connections and click 
to display the details of the connection.
Figure 216: SSH Control > Connections — Client side host key settings
-
Upload or paste the private part of the SSH host key.
SPS allows you to use the following SSH host keys:
-
RSA (ssh-rsa), which is the most widely used public-key algorithm for the SSH key. In SPS, uploading RSA SSH host keys are supported in PKCS #1 PEM, PKCS #8 PEM, OpenSSH (openssh-key-v1), and PuTTY private key formats.
NOTE:
One Identity recommends using 2048-bit RSA keys (or stronger).
-
Ed25519 (ssh-ed25519), which offers a better security and faster performance compared to RSA.
In SPS, uploading Ed25519 SSH host keys are supported in PKCS #8 PEM, OpenSSH (openssh-key-v1), and PuTTY private key formats.
-
ECDSA NIST P-256 (ecdsa-sha2-nistp256), which is a variant of the Digital Signature Algorithm (DSA). In SPS, uploading ECDSA SSH host keys are supported in SEC1 PEM, PKCS #8 PEM, OpenSSH (openssh-key-v1), and PuTTY private key formats.
You can have multiple SSH server host keys on SPS for the same server, however, you cannot set more than one key for each type. For example, you can keep your old RSA SSH key and generate a new Ed25519 key but you cannot set two RSA keys.
TIP: Click on the fingerprint to display the public part of the key.
Supported SSH channel types
The available SSH channel types and their functionalities are described below. For details on configuring Channel Policies, see Creating and editing channel policies. For a list of supported client applications, see Supported protocols and client applications.
-
Agent: Forwards the SSH authentication agent from the client to the server.
NOTE: To perform agent-based authentication on the target server, it is not required to enable the Agent-forwarding channel in the Channel Policy used by the connection. The Agent-forwarding channel is needed only to establish connections from the target server to other devices and authenticate using the agent running on the client.
-
X11 Forward: Forwards the graphical X-server session from the server to the client. Enter the address of the client into the Allow client address field to permit X11-forwarding only to the specified clients. Specify IP addresses or networks (in IP address/Prefix format).
NOTE: Certain client applications send the Target address as a hostname, while others as an IP address. If you are using a mix of different client applications, you might have to duplicate the channel rules and create IP-address and hostname versions of the same rule.
-
Local Forward: Forwards traffic arriving to a local port of the client to a remote host. To enable forwarding only between selected hosts, enter their IP addresses into the Details field. If the Details field is empty, local forwarding is enabled without restriction, the client may forward any traffic to the remote host. Enter the source of the forwarded traffic into the Originator, the target of the traffic into the Target field. Specify IP addresses or networks (in IP address/Prefix format). These parameters are the end-points of the forwarded traffic (that is, the local host that sends data to the remote host), and not the SSH server or the client.
For example, to enable forwarding from the 192.168.20.20 host to the remote host 192.168.50.50, enter 192.168.20.20 into the Originator, and 192.168.50.50 into the Target field.
Figure 217: Local TCP forwarding
NOTE: Certain client applications send the Originator and Target addresses as hostnames, while others as IP addresses. If you are using a mix of different client applications, you might have to duplicate the channel rules and create IP-address and hostname versions of the same rule.
|
|
Caution:
Port forwarding across One Identity Safeguard for Privileged Sessions (SPS) may fail for certain SSH client-server combinations. This happens if within the protocol, the address of the remote host is specified as a hostname during the port-forwarding request (SSH_MSG_GLOBAL_REQUEST), but the hostname is resolved to IP address in the channel opening request (SSH_MSG_CHANNEL_OPEN. By default, SPS rejects such connections.
To enable these connections, navigate to SSH Control > Settings, and disable the Strict mode option. |
-
Remote Forward: Forwards traffic arriving a remote port of the server to the client. To enable forwarding only between selected hosts, enter their IP addresses into the Details field. If the Details field is empty, remote forwarding is enabled without restriction, the SSH server may forward any traffic to the client. Enter the source of the forwarded traffic into the Originator, the target of the traffic into the Target field. Specify IP addresses or networks (in IP address/Prefix format). These parameters are the end-points of the forwarded traffic (that is, the remote host that sends data to the client), and not the SSH server.
For example, to enable forwarding from the 192.168.20.20 remote host to the client 192.168.50.50, enter 192.168.20.20 into the Originator, and 192.168.50.50 into the Target field.
Figure 218: Remote TCP forwarding
NOTE: Certain client applications send the Originator and Target addresses as hostnames, while others as IP addresses. If you are using a mix of different client applications, you might have to duplicate the channel rules and create IP-address and hostname versions of the same rule.
|
|
Caution:
Port forwarding across SPS may fail for certain SSH client-server combinations. This happens if within the protocol, the address of the remote host is specified as a hostname during the port-forwarding request (SSH_MSG_GLOBAL_REQUEST), but the hostname is resolved to IP address in the channel opening request (SSH_MSG_CHANNEL_OPEN. By default, SPS rejects such connections.
To enable these connections, navigate to SSH Control > Settings, and disable the Strict mode option. |
-
Session Exec: Execute a remote command (for example rsync) without opening a session shell. Enter the permitted command into the Permitted commands field. You can use regular expressions to specify the commands. This field can contain only letters (a-z, A-Z), numbers (0-9), and the following special characters ({}()*?\\|[]).
|
|
Caution:
Restricting the commands available in Session Exec channels does not guarantee that no other commands can be executed. Commands can be renamed, or executed from shell scripts to circumvent such restrictions. |
-
Session Exec SCP: Transfers files using the Secure Copy (SCP) protocol.
-
To make the list of file operations available in the File operations column of the Search page, navigate to the Channel Policies page of the protocol, and enable the Log file transfers to database option. This option is disabled by default.
-
To send the file operations into the system log, enable the Log file transfers to syslog option. This option is disabled by default.
NOTE: Turning logging on might result in a slight performance penalty. If traffic load slows processes down, disable the option.
|
|
Caution:
The WinSCP application does not follow the RFC of the SCP protocol properly, but transfers files in a Session Shell channel instead of a Session Exec SCP channel. This has the following results:
-
If the Session Shell channel is enabled in a Channel Policy (this is needed for SSH terminal sessions as well), and your users use WinSCP using the File protocol > SCP option, they will be able to transfer files to and from the server. Also, these files will not be listed in the File operations field of the Search page.
-
To avoid these problems, you have to enforce that your clients use WinSCP with the File protocol > SFTP option. WinSCP uses SFTP by default, but can be changed manually to use SCP, and also falls back to using SCP if a server rejects SFTP.
-
To terminate the connection when a user transfers a file with WinSCP using the Session Shell channel, create a Content Policy that matches the WinSCP: this is end-of-file string in screen content, and use this policy in your Connection Policies. For details on Content Policies, see Real-time content monitoring with Content Policies. This solution has been tested with WinSCP version 5.1.5: if it does not work for your version, contact our Support Team. |
-
Session Subsystem: Use a subsystem. Enter the name of the permitted subsystem into the Permitted subsystem field.
-
Session SFTP: Transfers files using the Secure File Transfer Protocol (SFTP).
-
To make the list of file operations available in the File operations column of the Search page, navigate to the Channel Policies page of the protocol, and enable the Log file transfers to database option. This option is disabled by default.
-
To send the file operations into the system log, enable the Log file transfers to syslog option. This option is disabled by default.
NOTE: Turning logging on might result in a slight performance penalty. If traffic load slows processes down, disable the option.
-
Session Shell: The traditional remote terminal session.
|
|
Caution:
The WinSCP application does not follow the RFC of the SCP protocol properly, but transfers files in a Session Shell channel instead of a Session Exec SCP channel. This has the following results:
-
If the Session Shell channel is enabled in a Channel Policy (this is needed for SSH terminal sessions as well), and your users use WinSCP using the File protocol > SCP option, they will be able to transfer files to and from the server. Also, these files will not be listed in the File operations field of the Search page.
-
To avoid these problems, you have to enforce that your clients use WinSCP with the File protocol > SFTP option. WinSCP uses SFTP by default, but can be changed manually to use SCP, and also falls back to using SCP if a server rejects SFTP.
-
To terminate the connection when a user transfers a file with WinSCP using the Session Shell channel, create a Content Policy that matches the WinSCP: this is end-of-file string in screen content, and use this policy in your Connection Policies. For details on Content Policies, see Real-time content monitoring with Content Policies. This solution has been tested with WinSCP version 5.1.5: if it does not work for your version, contact our Support Team. |
Authentication Policies
An authentication policy is a list of authentication methods that can be used in a connection. Connection definitions refer to an authentication policy to determine how the client can authenticate to the target server. Separate authentication methods can be used on the client and the server-side of the connection.
Figure 219: Authentication policies
Topics:
Creating a new authentication policy
The following describes how to create a new authentication policy.
To create a new authentication policy
-
Navigate to SSH Control > Authentication Policies, and click 
.
Figure 220: SSH Control > Authentication Policies — Configuring authentication policies
-
Enter a name for the policy.
-
Select the gateway authentication method to allow authentication on SPS. Note that this is an inband authentication that happens within the SSH protocol.
If you selected Kerberos as the gateway authentication method, SPS disables all other authentication backends and authentication methods.
-
Select the authentication database used on the client-side in the Authentication backend field. For details on the client-side authentication settings, see Client-side authentication settings.
If you selected Kerberos as the gateway authentication method, skip this step. For details, see Kerberos authentication settings.
-
Select the authentication method used on the server-side in the Relayed authentication methods field. For details on the relayed authentication settings, see Relayed authentication methods. Note the following:
-
If you selected Kerberos as the gateway authentication method, skip this step. For details, see Kerberos authentication settings.
-
If you selected Public key > Agent as the relayed authentication method:
If this option is used, SPS requests the client to use its SSH agent to authenticate on the target server. Therefore, you must configure your clients to enable agent forwarding, otherwise authentication will fail. For details on enabling agent forwarding in your SSH application, see the documentation of the application.
To be able to select Kerberos as the relayed authentication method, ensure that you also select Kerberos as the gateway authentication method.
-
Click 
.
NOTE: Consider the following:
-
The client-side authentication settings apply for authenticating the user inband (that is, within the SSH protocol) to the One Identity Safeguard for Privileged Sessions (SPS) gateway, and is independent from the gateway authentication performed on the SPS web interface. The web-based gateway authentication is an out-of-band gateway authentication method that can be required by the connection policy. For details on out-of-band gateway authentication, see Configuring out-of-band gateway authentication.
Gateway authentication on the SPS web interface can be used together with authentication policies. In an extreme setting, this would mean that the user has to perform three authentications: a client-side gateway authentication within the SSH protocol to SPS, an out-of-band gateway authentication on the SPS web interface, and a final authentication on the target server.
-
The Connection Policy will ignore the settings for server-side authentication (set under Relayed authentication methods for SSH protocol) if a Credential Store is used in the Connection Policy.
