Locating the user

No matter how the LDAP module is called (via its authorize, authenticate, accounting methods or the %ldap.memberof() xlat) the first operation the module performs it to populate &control.LDAP-UserDN with the location of the authenticating user’s object in LDAP.

Below is an example of configuring a user section for OpenLDAP, with callouts explaining how the different configuration items are used. See ldapsearch for how to determine the appropriate values for your directory.

Editing mods-available/ldap to specify how to search for user objects

ldap {
	user {
		# example - ou=people,${..base_dn}
		base_dn = "<user.base_dn>" (1)

		# example - (&(objectClass=posixAccount)(uid=%{&Stripped-User-Name || &User-Name}))
		filter = "(&(<user.filter>)(<user_uid_attribute>=%{&Stripped-User-Name || &User-Name}))"  (2)
	}
}

1	Where in the directory to begin the search for the user object. This must be a point in the directory tree containing the objects for all users that might authenticate.
2	The filter used to locate the user object belonging to the authenticating user. The user filter is normally a combination of two subfilters. The first subfilter specifies the objectClass of user objects. The second filter selects a user object matching the username the user provided at login.

It is important that the user.base_dn and user.filter values are restrictive enough to return only a single entry when searching for users. If multiple entries are returned the LDAP module will fail.

See User object disambiguation for strategies for limiting the number of user object entries returned.