Knox Gateway Administration Guide
Also available as:
PDF
loading table of contents...

Chapter 3. Defining Cluster Topologies

The Knox Gateway supports one or more Hadoop clusters. Each Hadoop cluster configuration is defined in a topology deployment descriptor file in the $gateway/conf/topologies directory and is deployed to a corresponding WAR file in the $gateway/data/deployments directory. These files define how the gateway communicates with each Hadoop cluster.

The descriptor is an XML file contains the following sections:

  • gateway/provider -- configuration settings enforced by the Knox Gateway while providing access to the Hadoop cluster.

  • service -- defines the Hadoop service URLs used by the gateway to proxy communications from external clients.

The gateway automatically redeploys the cluster whenever it detects a new topology descriptor file, or detects a change in an existing topology descriptor file.

The following table provides an overview of the providers and services:

Table 3.1. Cluster Topology Provider and Service Roles

TypeRoleDescription
gateway/providerhostmapMaps external to internal node hostnames, replacing the internal hostname with the mapped external name when the hostname is embedded in a repsonse from the cluster.
 authenticationIntegrates an LDAP store to authenticate external requests accessing the cluster via the Knox Gateway. Refer to Set Up LDAP Authentication for more information.
 federationDefines HTTP header authentication fields for an SSO or federation solution provider. Refer to Set up HTTP Header Authentication for Federation/SSO
 identity-assertionResponsible for the way that the authenticated user's identity is asserted to the service that the request is intended for. Also maps external authenticated users to an internal cluster that the gateway asserts as the current session user or group. Refer to Configure Identity Assertion for more information.
 authorizationService level authorization that restricts cluster access to specified users, groups, and/or IP addresses. Refer to Configure Service Level Authorization for more information.
 webappspecConfigures a web application security plugin that provides protection filtering against Cross Site Request Forgery Attacks. Refer to Configure Web Application Security for more information.
HA providerhigh availabilitySyncs all Knox instances to use the same topologies credentials keystores.
service$service_nameBinds a Hadoop service with an internal URL that the gateway uses to proxy requests from external clients to the internal cluster services. Refer to Configure Hadoop Service URLs for more information.

Cluster topology descriptors have the following XML format:

<topology>
 <gateway>
 <provider>
 <role> </role>
 <name> </name>
 <enabled> </enabled>
 <param>
 <name> </name>
 <value> </value>
 </param>
 </provider>
 </gateway>
 <service> </service>
</topology>