Using LDAP authentication for Presto on Amazon EMR

Follow the steps in this section to configure LDAP. See each step for examples and links to more information.

Step 1: Gather information about your LDAP server and copy the server certificate to Amazon S3

You'll need the information and items in the following section from your LDAP server to configure LDAP authentication.

The IP address or host name of the LDAP server

The Presto coordinator on the Amazon EMR master node must be able to reach the LDAP server at the specified IP address or host name. By default, Presto communicates with the LDAP server using LDAPS over port 636. If your LDAP implementation requires a custom port, you can specify it using the ldap.url property with Amazon EMR 5.16.0 or later, or using authentication.ldap.url with earlier versions. Substitute the custom port for 636 as shown in the presto-config configuration classification examples in Step 3: Create a configuration JSON with Presto properties for LDAP. Ensure that any firewalls and security groups allow inbound and outbound traffic on port 636 (or your custom port) and also port 8446 (or your custom port), which is used for internal cluster communications.

The LDAP server certificate

You must upload the certificate file to a secure location in Amazon S3. For more information, see How do I upload files and folders to an S3 Bucket in the Amazon Simple Storage Service User Guide. You create a bootstrap action that copies this certificate from Amazon S3 to each node in the cluster when the cluster launches. In Step 4: Create the script to copy the LDAP server certificate and upload it to Amazon S3. The example certificate is s3://amzn-s3-demo-bucket/ldap_server.crt.

The LDAP server's settings for anonymous binding

If anonymous binding is disabled on PrestoDB, you need the user ID (UID) and password of an account with permissions to bind to the LDAP server so that the PrestoDB server can establish a connection. You specify the UID and password using the internal-communication.authentication.ldap.user and internal-communication.authentication.ldap.password properties in the presto-config configuration classification. Amazon EMR 5.10.0 does not support these settings, so anonymous binding must be supported on the LDAP server when you use this release version.

Note that Trino doesn't require the anonymous binding configuration.

The example configurations in Step 3: Create a configuration JSON with Presto properties for LDAP include this account for clarity, with the exception of the 5.10.0 example, where it is not supported. If the LDAP server uses anonymous binding, remove the internal-communication.authentication.ldap.user and internal-communication.authentication.ldap.password name/value pairs.

The LDAP distinguished name (DN) for Presto users

When you specify the LDAP configuration for Presto, you specify a bind pattern that consists of ${USER} along with an organizational unit (OU) and additional domain components (DCs). Presto replaces ${USER} with the actual User ID (UID) of each user during password authentication to match the distinguished name (DN) that this bind pattern specifies. You need the OUs that eligible users belong to and their DCs. For example, to allow users from the admins OU in the corp.example.com domain to authenticate to Presto, you specify ${USER},ou=admins,dc=corp,dc=example,dc=com as the user bind pattern.

When using Amazon EMR 5.10.0, you can specify only one such pattern. Using Amazon EMR 5.11.0 or later, you can specify multiple patterns separated by a colon (:). Users attempting to authenticate to Presto are compared to the first pattern, then the second, and so on. For an example, see Step 3: Create a configuration JSON with Presto properties for LDAP.

Step 2: Set up a security configuration

Create a security configuration with in-transit encryption enabled. For more information, see Create a security configuration in the Amazon EMR Management Guide. The encryption artifacts that you provide when you set up in-transit encryption are used to encrypt internal communication between Presto nodes. For more information, see Providing certificates for in-transit data encryption. The LDAP server certificate is used to authenticate client connections to the Presto server.

Step 3: Create a configuration JSON with Presto properties for LDAP

You use the presto-config configuration classification to set Presto properties for LDAP. The format and contents of presto-config are slightly different depending on the Amazon EMR release version and the Presto installation (PrestoDB or Trino). Examples of configuration differences are provided later in this section. For more information, see Configure applications.

The following steps assume that you save the JSON data to a file, MyPrestoConfig.json. If you use the console, upload the file to a secure location in Amazon S3 so that you can reference it when you create the cluster. If you use the AWS CLI, you can reference the file locally.

[
   {
      "Classification":"prestosql-config",
      "Properties":{
         "http-server.authentication.type":"PASSWORD"
      }
   },
   {
      "Classification":"prestosql-password-authenticator",
      "Properties":{
         "password-authenticator.name":"ldap",
         "ldap.url":"ldaps://ip-xxx-xxx-xxx-xxx.ec2.internal:636",
         "ldap.user-bind-pattern": "uid=${USER},ou=admins,dc=ec2,dc=internal:uid=${USER},ou=datascientists,dc=ec2,dc=internal"
      }
   }
]

[{
        "Classification": "presto-config",
                "Properties": {
                        "http-server.authentication.type": "PASSWORD"
                }
        },
        {
                "Classification": "presto-password-authenticator",
                "Properties": {
                        "password-authenticator.name": "ldap",
                        "ldap.url": "ldaps://ip-xxx-xxx-xxx-xxx.ec2.internal:636",
                        "ldap.user-bind-pattern": "uid=${USER},ou=admins,dc=ec2,dc=internal:uid=${USER},ou=datascientists,dc=ec2,dc=internal",
                        "internal-communication.authentication.ldap.user": "presto",
                        "internal-communication.authentication.ldap.password": "123456"
                }
        }]

[{
        "Classification": "presto-config",
                "Properties": {
                        "http-server.authentication.type": "LDAP",
                        "authentication.ldap.url": "ldaps://ip-xxx-xxx-xxx-xxx.ec2.internal:636",
                        "authentication.ldap.user-bind-pattern": "uid=${USER},ou=admins,dc=ec2,dc=internal:uid=${USER},ou=datascientists,dc=ec2,dc=internal",
                        "internal-communication.authentication.ldap.user": "presto",
                        "internal-communication.authentication.ldap.password": "123456"
                }
        }]

[{
        "Classification": "presto-config",
                "Properties": {
                        "http-server.authentication.type": "LDAP",
                        "authentication.ldap.url": "ldaps://ip-xxx-xxx-xxx-xxx.ec2.internal:636",
                        "ldap.user-bind-pattern": "uid=${USER},ou=prestousers,dc=ec2,dc=internal"
                }
        }]

Step 4: Create the script to copy the LDAP server certificate and upload it to Amazon S3

Create a script that copies the certificate file to each node in the cluster and adds it to the keystore. Create the script using a text editor, save it, and then upload it to Amazon S3. In Step 5: Create the cluster, the script file is referenced as s3://amzn-s3-demo-bucket/LoadLDAPCert.sh.

The following example script uses the default keystore password, changeit. We recommend that you connect to the master node after you create the cluster and change the keystore password using the keytool command.

#!/bin/bash
aws s3 cp s3://amzn-s3-demo-bucket/ldap_server.crt .
sudo keytool -import -keystore /usr/lib/jvm/jre-1.8.0-openjdk.x86_64/lib/security/cacerts -trustcacerts -alias ldap_server -file ./ldap_server.crt -storepass changeit -noprompt

Step 5: Create the cluster

When you create the cluster, you specify Presto and other applications that you want Amazon EMR to install. The following examples also reference the configuration classification properties within a JSON, but you can also specify the configuration classification inline.