The Hadoop authentication plugin enables Solr to use the Hadoop authentication library for securing Solr nodes.
This authentication plugin is a thin wrapper that delegates all functionality to the Hadoop authentication library. All configuration parameters for the library are passed through the plugin.
This plugin can be particularly useful in leveraging an extended set of features or newly available features in the Hadoop authentication library.
Please note that the version of Hadoop library used by Solr is upgraded periodically. While Solr will ensure the stability and backwards compatibility of the structure of the plugin configuration (viz., the parameter names of this plugin), the values of these parameters may change based on the version of Hadoop library. Please review the Hadoop documentation for the version used by your Solr installation for more details.
For some of the authentication schemes (e.g., Kerberos), Solr provides a native implementation of authentication plugin. If you require a more stable setup, in terms of configuration, ability to perform rolling upgrades, backward compatibility, etc., you should consider using such plugin. Please review the section Authentication and Authorization Plugins for an overview of authentication plugin options in Solr.
There are two plugin classes:
-
HadoopAuthPlugin
: This can be used with standalone Solr as well as Solrcloud with PKI authentication for internode communication. -
ConfigurableInternodeAuthHadoopPlugin
: This is an extension of HadoopAuthPlugin that allows you to configure the authentication scheme for internode communication.
For most SolrCloud or standalone Solr setups, the |
Plugin Configuration
class
-
Should be either
solr.HadoopAuthPlugin
orsolr.ConfigurableInternodeAuthHadoopPlugin
. This parameter is required. type
-
The type of authentication scheme to be configured. See configuration options. This parameter is required.
sysPropPrefix
-
The prefix to be used to define the Java system property for configuring the authentication mechanism. This property is required.
The name of the Java system property is defined by appending the configuration parameter name to this prefix value. For example, if the prefix is
solr
then the Java system propertysolr.kerberos.principal
defines the value of configuration parameterkerberos.principal
. authConfigs
-
Configuration parameters required by the authentication scheme defined by the type property. This property is required. For more details, see Hadoop configuration options.
defaultConfigs
-
Default values for the configuration parameters specified by the
authConfigs
property. The default values are specified as a collection of key-value pairs (i.e.,"property-name": "default_value"
). enableDelegationToken
-
If
true
, the delegation tokens functionality will be enabled. initKerberosZk
-
For enabling initialization of kerberos before connecting to ZooKeeper (if applicable).
proxyUserConfigs
-
Configures proxy users for the underlying Hadoop authentication mechanism. This configuration is expressed as a collection of key-value pairs (i.e.,
"property-name": "default_value"
). clientBuilderFactory
-
No | The
HttpClientBuilderFactory
implementation used for the Solr internal communication. Only applicable forConfigurableInternodeAuthHadoopPlugin
.
Example Configurations
Kerberos Authentication using Hadoop Authentication Plugin
This example lets you configure Solr to use Kerberos Authentication, similar to how you would use the Kerberos Authentication Plugin.
After consulting the Hadoop authentication library’s documentation, you can supply per host configuration parameters using the solr.*
prefix. As an example, the Hadoop authentication library expects a parameter kerberos.principal
, which can be supplied as a system property named solr.kerberos.principal
when starting a Solr node. Refer to the section Kerberos Authentication Plugin for other typical configuration parameters.
Please note that this example uses ConfigurableInternodeAuthHadoopPlugin
, and hence you must provide the clientBuilderFactory
implementation. As a result, all internode communication will use the Kerberos mechanism, instead of PKI authentication.
To setup this plugin, use the following in your security.json
file.
{
"authentication": {
"class": "solr.ConfigurableInternodeAuthHadoopPlugin",
"sysPropPrefix": "solr.",
"type": "kerberos",
"clientBuilderFactory": "org.apache.solr.client.solrj.impl.Krb5HttpClientBuilder",
"initKerberosZk": "true",
"authConfigs": [
"kerberos.principal",
"kerberos.keytab",
"kerberos.name.rules"
],
"defaultConfigs": {
}
}
}
Simple Authentication with Delegation Tokens
Similar to the previous example, this is an example of setting up a Solr cluster that uses delegation tokens. Refer to the parameters in the Hadoop authentication library’s documentation or refer to the section Kerberos Authentication Plugin for further details. Please note that this example does not use Kerberos and the requests made to Solr must contain valid delegation tokens.
To setup this plugin, use the following in your security.json
file.
{
"authentication": {
"class": "solr.HadoopAuthPlugin",
"sysPropPrefix": "solr.",
"type": "simple",
"enableDelegationToken":"true",
"authConfigs": [
"delegation-token.token-kind",
"delegation-token.update-interval.sec",
"delegation-token.max-lifetime.sec",
"delegation-token.renewal-interval.sec",
"delegation-token.removal-scan-interval.sec",
"cookie.domain",
"signer.secret.provider",
"zk-dt-secret-manager.enable",
"zk-dt-secret-manager.znodeWorkingPath",
"signer.secret.provider.zookeeper.path"
],
"defaultConfigs": {
"delegation-token.token-kind": "solr-dt",
"signer.secret.provider": "zookeeper",
"zk-dt-secret-manager.enable": "true",
"token.validity": "36000",
"zk-dt-secret-manager.znodeWorkingPath": "solr/security/zkdtsm",
"signer.secret.provider.zookeeper.path": "/token",
"cookie.domain": "127.0.0.1"
}
}
}