Multiple Distributed LDX instances in a Modified Federated Topology

In some highly active environments, a single LDX server in the modified federated topology may not be sufficient. To alleviate this, one or more LDX instances can be deployed to manage versioned links for individual business units that only index the application sources used by that business unit.

With multiple LDXs, it is typical to configure each to index the same set of data sources as the original single LDX topology

• Each instance can be made to handle the traffic of any other LDX instance should that particular LDX system be down or undergo maintenance (e.g. Backup, compaction, validation)
• Each instance’s backup can be restored on other LDX instances as the data they index and their context root is the same
• If clustered via HAProxy, you can provide load balancing with the multiple LQE’s supporting requests

An alternative topology if a business unit has a very narrow focus for which links to other artifacts are very strictly constrained, would be to index the needed subset of data sources to fulfill queries from the BU specific LDX. Caution: if a user creates a link to an artifact outside the constraint, the link will not be seen by other users and eventually will not be seen by the user creating the link (temporary cache expires), the tool where the link is stored should be added to the particular LDX instance so that those links can be indexed and shown to the end user in the expanded scope.

Multiple Distributed LDX Servers in Modified Federated Topology

This document will provide the steps to configure a Modified Federated Topology to include additional LDX servers which are scoped for a particular business unit.

Deploy additional LDX server

• The additional LDX application should be deployed on its own server
• Install LDX application per the documentation
  • If Jazz Authorization Server is configured in the environment, make sure to check the box in the install to enable Jazz SSO during install.

Configure additional LDX server for the business unit

• A restore using the backup of an existing LDX server can be used to make a clone.

If an LDX instance will be used only for a single narrowly scoped business unit, a custom configuration is required.

• Configure LDX data sources
  • The configured data sources may include all data sources.

In both configurations, the application servers will need to point to an LDX server:

• Configure DN and ETM advanced properties to point to an LDX server.
  • Change the value in the server Advanced Properties->com.ibm.team.links.service.internal.LinkIndexService->Link index provider URL as needed

Validation and testing the configuration

• Once configured the updated RM applications will only make link lookup requests to the LDX server configured in Advanced Properties.
• The new LDX server(s) will only have data sources for the RM, QM, and CCM instances that have links to or from the business unit RM server.
  • A link to or from the RM server to another application server that is not indexed by the business unit LDX server will not be shown in an RM artifact.
  • Links to or from the RM server from the application servers that are data sources in the business unit LDX server will be shown in RM artifacts.
  • If all data sources are present in all LDX instances, the previous points will not apply.

Related topics: [StandardTopologyPatternsOverview][StandardTopologyPatternsOverview]], DeploymentPlanningAndDesign