Create an initial or new Search indexEnable Search functionality by building the index and copying it to each node that is running the Search application.
Before you begin
If you are installing a non-English language deployment and you are creating a Search index for the first time, enable the relevant language dictionaries for your deployment before creating the index. For more information, see Configure dictionaries for Search.
About this task
In IBM Connections 3 and later, each node running the Search application must have the Search index stored locally on the node's file system. This configuration change is required due to limitations with Apache Lucene indexes that are accessed over network shares. Because there are multiple indexes in a clustered environment, they must all be kept in synchronization with each other. During update indexing, documents are first written to a cache table in the HOMEPAGE database and then written to each index across the nodes. When a new index needs to be built, the database cache is skipped, and the crawling and indexing process writes directly to the index directory on the node that is performing the indexing task. After indexing has been completed, the complete index must be copied to the other nodes in the cluster that are running Search. Search functionality is not available until the index is built and deployed to each node. The index is built and ready to be copied when the INDEX.READY and CRAWLING_VERSION files are present in the index directory.
Note: Initial index creation occurs when any scheduled indexing task fires and an index does not yet exist.
- Stop all the nodes that are running the Search application. If there are existing search indexes on these nodes, delete them by performing the steps described in Delete the index.
- Start one Search node in the cluster.
- Wait for the index to be built. You can tell that the index is built when the INDEX.READY and CRAWLING_VERSION files are present in the index directory.
If you are running Search on a single node only, the task is now complete; you do not need to follow any more of the steps listed. If you are running multiple Search nodes in the cluster, complete the rest of the steps.
- Stop the Search node on which the index was built.
- Copy the built index directory to the index directory on each node that will be running the Search application.
- Start all the nodes.
Parent topicAdminister Search
Configure dictionaries for Search
Install IBM Connections 3.0.1
Delete the indexVerifying Search Related references
Search default indexing tasks
April 11, 2011 3:29:15 AM
Apr 11, 2011 3:29:15 AM 3 Apr 8, 2011 8:45:12 AM 2 Apr 8, 2011 8:16:50 AM 1