When you install HCL Connections,
Search indexing is automatically configured. To create the initial
Search index, all you need to do is wait for one of the default indexing
tasks to run.
About this task
Initial index creation occurs when any scheduled indexing task fires and an index does not
yet exist. As part of the initial index creation process, the index is automatically rolled out
to the secondary nodes in your deployment. Each node running the Search application must have the
Search index stored locally on the node's file system. Because there are multiple indexes in a
clustered environment, they must all be kept in synchronization with each other. The Search
index directory is defined by the IBM®
WebSphere® Application Server variable
SEARCH_INDEX_DIR. You can change the location of the index by editing this variable. For more
information, see Changing the location of the Search index.
After the initial
index has been built and optimized, the contents of the index directory are copied to a staging
folder. When the newly-built index is successfully posted, JMS messages are broadcast so that
each node automatically downloads the index from the staging folder and loads it. The index
management tables are populated at the same time. For Search to function properly, the initial
index must have completed successfully and it must be deployed to all nodes.
Important: Do not stop your deployment until the index has been copied to all nodes. If
the server is stopped during this process, the index will not be successfully rolled out to all
nodes. In this event, you need to manually copy the index from the staging location to the
other nodes.
You can change the location of the Search index staging folder by
editing the WebSphere® Application Server variable,
SEARCH_INDEX_SHARED_COPY_LOCATION. If you do so, ensure that the staging folder is accessible to
all nodes in the cluster and that the filesystem permissions on the folder are sufficient; the
websphere user must have read/write permissions on this location.