Configuring Helix Authentication Service

The authentication service is configured using environment variables. Because there are numerous settings, it is easiest to create a file called .env that contains all of the settings. If you change the .env file while the service is running, the service must be restarted for the changes to take effect.

The choice of process manager affects how these environment variables are defined. For further details, see Starting Helix Authentication Service.

The example.env and .env files

We recommend that you use the example.env file as your starting point for configuring the service. The example.env file is in your installation directory. The example.env file is a plain text file that contains:

  • NAME=VALUE pairs

  • comment lines starting with #

For example:

# this is a comment

If you use example.env as your starting point, rename it to .env because the service needs that name to operate.

Make sure your .env file is located in the current working directory when the service is started. Typically this is the same directory as the package.json file of the service code.


Be aware that the configuration of the service with your identity provider might require some assistance from Perforce Support.


We recommend that you use the script because it is easier than a manual configuration of .env file.

The configuration script is in the bin directory of your installation. For example, if you installed HAS using the package, to get the help for the configuration script, type

/opt/perforce/helix-auth-svc/bin/ -h

and the output will be:



We strongly recommend that you use proper certificates and a trusted certificate authority (CA). A self-signed certificate might be rejected at any time.

The Helix Authentication Service reads its certificate and key files using the paths defined in CERT_FILE and KEY_FILE, respectively. The path for the CA certificate is read from the CA_CERT_FILE environment variable. Providing a CA file is only necessary if the CA is not one of the root authorities whose certificates are already installed on the system. Clients accessing the /requests/status/:id route will require a valid client certificate signed by the certificate authority.


The 2021.2 release and later also support the use of a PFX_FILE. If a PFX_FILE is given, CERT_FILE and KEY_FILE are ignored. See Other Settings.

In 2022.1 and later, CERT_FILE and KEY_FILE are the aliases we recommend you use because the old names, SP_CERT_FILE and SP_KEY_FILE, have been deprecated.

If the certificate files are changed, the service will need to be restarted because the service only reads the files at startup.

Restarting the service

If you installed by using the packager or the installation script, HAS is already running and its helix-auth is its service name. However, any changes to the HAS environment settings only take effect when the service is restarted.

  • If you installed the HAS by using the package or install script, use the systemd system manager to restart the helix-auth service.
  • If you are running the service under Windows, see Installing as a Windows service .

OpenID Connect settings variables

Variable Name Description
OIDC_CLIENT_ID The client identifier as provided by the OIDC identity provider
OIDC_CLIENT_SECRET The client secret as provided by the OIDC identity provider
OIDC_CLIENT_SECRET_FILE Path of the file containing the client secret as provided by the OIDC identity provider. This setting should be preferred over OIDC_CLIENT_SECRET to prevent accidental exposure of the client secret.

The default behavior is to detect the supported methods in the OIDC issuer data. Therefore, in most cases it is optional to specify the authorization code challenge method, which is either S256 or plain.

However, not all identity providers supply this information, in which case this setting becomes necessary.

OIDC_ISSUER_URI The OIDC provider issuer URL

To prompt the user to select an account when authenticating with an OIDC identity provider, set to true. This might be useful if the IdP supports the user switching between accounts, such as personal account and work account.

OpenID Connect has a discovery feature in which the identity provider advertises various properties via a "discovery document". The URI path will be /.well-known/openid-configuration at the IdP base URL, which is described in the OpenID Connect (OIDC) specification. This information makes the process of configuring an OIDC client easier.

The OIDC client identifier and secret are generally provided by the identity provider during the setup and configuration of the application, and this is unique to each identity provider. For guidance with several popular identity providers, see Example Identity Provider configurations.

The OIDC issuer URI value is advertised in the discovery document mentioned above, and will be a property named issuer. Copy this value to the OIDC_ISSUER_URI service setting. Do NOT to use one of the "endpoint" URL values, because those are different from the issuer URI.

SAML settings variables


SAML identity providers advertise some of the information for the variables through their metadata URL (or metadata file).

The URL is different for each provider, unlike OIDC. See Example Identity Provider configurations.

When configuring the service as a "service provider" within a SAML identity provider, provide an entityID that is unique within your set of registered applications. By default, the service uses the value, which can be changed by setting the SAML_SP_ENTITY_ID environment variable. Anywhere that appears in this documentation, replace it with the value you defined in the identity provider.



Path of the file containing the public certificate of the identity provider, which is used to validate signatures of incoming SAML responses.


For security, we strongly recommend setting IDP_CERT_FILE. Otherwise, an attacker might try to pose as the IdP and forge SAML assertions.


Setting IDP_CERT_FILE is necessary when:


  • The metadata provided by the IdP lacks the public certificate of the IdP.

Default none


Description Path of the configuration file that defines SAML service providers that will be connecting to the authentication service.

When the authentication service is acting as a SAML identity provider, it reads some of its settings from a configuration file in the auth service installation. By default, this file is named saml_idp.conf.cjs and is identified by the IDP_CONFIG_FILE environment variable. It is evaluated using the Node.js require() function.



The authn context defines the method by which the user will authenticate with the IdP. Normally the default value works on most systems, but it may be necessary to change this value. For example, Azure might want this set to urn:oasis:names:tc:SAML:2.0:ac:classes:Password in certain cases.

The SAML_AUTHN_CONTEXT setting can be either a single value or multiple values.

If you specify multiple values, enclose them as a comma-separated list within square brackets. The entire value of the .env configuration file must be on one line. For example:

[urn:oasis:names:tc:SAML:2.0:ac:classes:Password, urn:oasis:names:tc:SAML:2.0:ac:classes:PasswordProtectedTransport]

Spaces and quotes are optional.





If set to true, do not request a specific authentication context from the SAML IdP. Instead, allow Active Directory Federation Services (ADFS) to decide. Some configurations use different contexts depending on how the user is authenticating. For example, accessing the server from outside the corporate network might correspond to a different context than accessing the server from inside the corporate network.




Description The entity identifier for the identity provider. This is not required, but if provided, then the IdP issuer will be validated for incoming logout requests/responses.
Default none



Specifies the path of a local file that contains the identity provider's metadata in XML format. An example of a SAML IdP that uses a file instead of a URL is Google Workspace (G Suite). If you use SAML_IDP_METADATA_FILE, it takes precedence over SAML_IDP_METADATA_URL.

Default none



URL of the IdP metadata configuration in XML format.


If you fetch the IdP metadata by the SAML_IDP_METADATA_URL setting, several other settings might be configured automatically by the service. The settings that are configured automatically depends on the information the IdP advertises via the metadata. Possibilities include:

In the unlikely scenario that the IdP returns data that needs to be modified, set the correct value in the appropriate environment variable, such as SAML_NAMEID_FORMAT


Be aware that a more secure option is to set the individual SAML IdP values: SAML_IDP_SSO_URL, SAML_IDP_ENTITY_ID, and IDP_CERT_FILE instead of setting SAML_IDP_METADATA_URL. Although using the metadata URL is more convenient, it relies heavily on the HTTPS protocol and the IdP maintaining strict control over their domain name at all times.




Description URL of IdP Single Log-Out service.



Description URL of SSO IdP provider to which the user is redirected to establish authentication.



Description Name of the property in the user profile to be used if nameID is missing, which is likely to be the case when using another authentication protocol (such as OIDC) with the identity provider (such as Okta).


Note: Changing the configuration file requires restarting the service because Node caches the file contents in memory.



Specifies the SAML name identifier format for the IdP to return during a successful authentication request. The default value allows the IdP to choose the value. However, you might need to specify the format to avoid getting different results for different users.

A typical value is urn:oasis:names:tc:SAML:1.1:nameid-format:emailAddress

For the list of possible values, see section 8.3 of the SAML Core specification.




Description Service provider audience value for AudienceRestriction assertions.




The entity identifier (entityID) for the Helix Authentication Service.


Settings for JSON web token authentication

If you have fully- or semi-automated users, such as build systems, they can authenticate with JSON web token authentication when they need to. This is an alternative to you setting up long-lived tickets to avoid the burden of authenticating frequently. In the case of Azure, JSON web token authentication has the additional advantage of being managed by the cloud service provider, making it both convenient and secure.

The /oauth/validate endpoint takes a JSON web token via the Authorization header and verifies it. To enable HAS to verify the authenticity of the token, set the following settings:

Name Description Default
OAUTH_ALGORITHM The algorithm used to cryptographically sign the JSON Web Token. This setting is optional. The list of supported values is found here: none
OAUTH_AUDIENCE The "audience" value that will be compared to the aud field of the JSON Web Token to ensure that the token is valid for this service. none
OAUTH_ISSUER The issuer that provided the JSON Web Token to the client, to be compared to the iss field of the JSON Web Token to ensure that it is from a trusted source. none
OAUTH_JWKS_URI The URI that provides the JSON Web Key Sets used to validate an JSON Web Token that was issued by the issuer identified in the OAUTH_ISSUER setting. none
OAUTH_TENANT_ID The "tenant identifier" that can be used to further restrict access to clients that have been granted permission to use this service. This setting is optional. none

JSON web token authentication involves the extension running in p4d, the client system by means of P4LOGINSSO, and the authentication service.

The loginhook extension checks the P4USER against a configured list of users and/or groups (configuration settings to be named client-sso-users and client-sso-groups), and indicates to p4d that the user must authenticate using the traditional SSO mechanism (using the P4LOGINSSO client-side program). When the client returns the token to p4d, the extension will send it to HAS to be verified. The returned payload must contain an identifier that can be mapped to the Helix Core Server user spec. Similar to web-based SSO, the administrator configures the name of the IdP property via the client-name-identifer setting, and the P4 user field is named via the client-user-identifiersetting, which can be user, email, or fullname.

The new settings client-name-identifer and client-user-identifier are necessary because the extension can still be used to authenticate other users via web-based SSO, which does not rely on P4LOGINSSO and web tokens, and thus will likely compare different sets of attributes than the "client-sso" users. For instance, web-based SSO users might use their email address, while client-sso users might use an "oid" that is compared to the user name.

Other Settings

Name Description Default
BIND_ADDRESS Define the IP address upon which the service will listen for requests. Setting this to (that is, localhost) means that only processes on the same host can connect, while a value of means requests made against any address assigned to the machine will work.
DEBUG Set to any value to enable debug logging in the service (writes to the console). none
CA_CERT_FILE Path of certificate authority file for service to use when verifying client certificates. none
CA_CERT_PATH Path of directory containing certificate authority files for service to use when verifying client certificates. All files in the named directory will be processed. none
CERT_FILE The service provider public certificate file, needed with SAML. none


By default, HAS accepts any valid client certificate that is signed by the designated certificate authority.


If you want additional security, consider using this setting or CLIENT_CERT_FP.

Specify a name or pattern to match against the Common Name in the client certificate used to acquire the user profile data. The patterns supported are described in the library at, with the asterisk (*) being the most common wildcard. For example:





This setting is even more secure than CLIENT_CERT_CN because the fingerprint is even more secure than the Common Name.

Can be set to the SHA256 digest of the client certificate (also called fingerprint) where a colon separates each byte. For example,


If set, the auth service compares this value to that of the client certificate used to retrieve the user profile data.

To generate the fingerprint, use the openssl command:

openssl x509 -noout -fingerprint -sha256 -in public-cert

DEFAULT_PROTOCOL The default authentication protocol to use. Can be oidc or saml. saml

If set to any non-empty value, will cause the service to require the user to authenticate, even if the user is already authenticated. For SAML, this means setting the forceAuthn field set to true, while for OIDC it will set the max_age parameter to 0. This is not supported by all identity providers, especially for OIDC.


If utmost security is a high priority, we recommend setting FORCE_AUTHN.

KEY_FILE The service provider private key file, typically needed with SAML. none
KEY_PASSPHRASE Passphrase to decrypt the private key of the PFX_FILE public/private key pair or the private key of KEY_FILE. The passphrase is provided as a string. For example,
KEY_PASSPHRASE_FILE This alternative to KEY_PASSPHRASE specifies the path to a UTF-8 encoded file that contains the passphrase to decrypt the private key of PFX_FILE or KEY_FILE. If KEY_PASSPHRASE_FILE is provided, KEY_PASSPHRASE is ignored. none
LOGGING Path of a logging configuration file. See the Logging section below. Setting this will override the DEBUG setting. none
LOGIN_TIMEOUT How long in seconds to wait for user to successfully authenticate before a timeout occurs in the login process. 60

These two settings:

  • specify the TLS client certificate to use when connecting to a Redis server using TLS

  • are only used if the REDIS_URL setting starts with rediss:// with a double ss instead of starting with redis:// with a single s

You can specify a value for these settings even if the Redis server does not currently use client certificates.


When connecting to Redis,

  • if this setting starts with rediss:// with a double ss, REDIS_CERT_FILE and REDIS_KEY_FILE are used and TLS is used to connect to Redis

  • if this setting starts with redis:// with a single letter s, no certificate-related settings are involved

PFX_FILE Path to a PKCS#12 formatted public/private key pair for TLS/SSL communications. Typically, the filename ends with .pfx and often PFX is used to refer to this file type. If the private key is encrypted, a passphrase must be provided to decrypt it. See KEY_PASSPHRASE and KEY_PASSPHRASE_FILE. none
SESSION_SECRET Password used for encrypting the in-memory session data. keyboard cat
SP_KEY_ALGO The algorithm used to sign the requests. sha256

The authentication service base URL visible to end users. Needs to match the application settings defined in IdP configuration.


If you use a load balancer in front of HAS, such as Amazon Web Services Elastic Load Balancing (ELB), the load balancer can have a certificate installed and use SSL termination (decryption). Such a process requires a protocol to forward the traffic to a port. Therefore, HAS supports setting the PORT and PROTOCOL configuration variables.

If SVC_BASE_URI is defined, it sets the value of PORT and PROTOCOL. For example, explicitly sets PROTOCOL to https and PORT to 443. In this scenario, 443 might also be considered the DNS name of the load balancer.

The default value of PORT is 3000, and the default value of PROTOCOL is http.

You can explicitly set PROTOCOL to http or https. The PORT value can be implicitly defined because http defaults to 80 and https defaults to 443.

Note that the PORT for SVC_BASE_URI is distinct from the syslog port described under Logging.


For information about configuring the service to treat the connection to the load balancer and/or proxy as secure, use the TRUST_PROXY setting. For details, see

SVC_GROUP The name of the group or user that runs the HAS process. HAS starts as the root user to bind to a port, then switches to the group user if SVC_GROUP is defined, or to the process owner if SVC_USER is defined. The package installation for HAS assigns the value of perforce to both settings in the .env file during installation. The values for SVC_USER and SVC_GROUP can be either names, such as perforce, or user identifiers, which are numbers. none


The authentication service will, by default, write only HTTP request logs to the console. With the DEBUG environment variable set to any value, additional logging will be written to the console. For more precise control, the LOGGING environment variable can be used to specify a logging configuration file. The format of the logging configuration file can be either JSON or JavaScript. Name the logging file with an extension of .json for a JSON file, or .cjs for a JavaScript file.

The top-level properties are listed in the table below.

Name Description Default
level Messages at this log level and above will be written to the named transport; follows syslog levels per RFC5424, section 6.2.1. Levels in order of priority: emerg, alert, crit, error, warning, notice, info, debug none
transport console, file, or syslog none

An example of logging messages to the console, starting explicitly at debug and including emerg:

module.exports = {
    level: 'debug',
    transport: 'console'

Logging to a named file can be achieved by setting the transport to file. Additional properties can then be defined within a property named file, as described in the table below.

Name Description Default
filename Path for the log file. auth-svc.log
maxsize Size in bytes before rotating the file. none
maxfiles Number of rotated files to retain. none

An example of logging messages to a named file, starting at the level of warning and including emerg, is shown below. This example also demonstrates log rotation by defining a maximum file size and a maximum number of files (maxfiles) to retain.

module.exports = {
    level: 'warning',
    transport: 'file',
    file: {
        filename: '/var/log/auth-svc.log',
        maxsize: 1048576,
        maxfiles: 10

Logging to the system logger, syslog, is configured by setting the transport to syslog. Additional properties can then be defined within a property named syslog, as described in the table below. Note that the syslog program name will be helix-auth-svc for messages emitted by the authentication service.

Name Description Default
facility Syslog facility, such as auth, cron, daemon, kern, mail, etc. local0
path Path of the syslog unix domain socket. For example, /dev/log none
host Host name of the syslog daemon. none
port Port number on which the syslog daemon is listening. none
protocol Communication protocol, such as tcp4, udp4, unix none

An example of logging all messages at levels from info up to emerg, to the system log, is shown below. This example demonstrates logging to syslog on Ubuntu 18, in which the default installation uses a unix domain socket that has /dev/log as its name.

module.exports = {
level: 'info',
transport: 'syslog',
syslog: {
path: '/dev/log',
protocol: 'unix'

Load configuration changes without restarting HAS

If you want to change the configuration for HAS, choose one of the following methods:

  • Edit the configuration file, stop the service instance, then restart it.

  • Signal the service process to reload the configuration.

    • One way is to use the Unix command kill -HUP <pid> to send a literal "signal" to the process. (This is similar to how some daemon programs might use SIGHUP as a signal to restart themselves to re-read a configuration file that has been changed.)

    • Another way is to send the USR2 signal to the process, get the process identifier of the service, and invoke the kill command:

      kill -USR2 <pid>

      An example of this command, and the correct process ID, are written to the log at info level when the service starts up.

      When HAS receives the USR2 signal, it reloads all the environment variables except those settings that cannot be changed during runtime:













See Starting Helix Authentication Service.