The default Lucene Index Server configuration is appropriate for a medium-sized Novell Teaming site. If you have a larger Teaming site, you can change its Lucene Index Server configuration.
After planning your Lucene configuration, complete the planning process for additional Advanced installation features as needed, then perform the Advanced installation as described in Section 10.0, Performing an Advanced Teaming Installation.
The Lucene Index Server is responsible for indexing all data on the Novell Teaming site so that Teaming users can easily use the Find and Search features to retrieve the information that they need. Text posted in folder entries is easy to index, because the formatting is simple. However, text in attached files arrives in many different file formats, many of which require conversion before the text in the files can be indexed. Therefore, the Lucene Index Server is dependent on the available file conversion technology in order to perform its indexing function. For information about the file viewers that Teaming uses, see Section 2.3.1, File Viewer Support.
The Lucene Index Server provides additional services on your Teaming site in addition to indexing. In fact, you cannot access your Teaming site if the Lucene Index Server is not running. For this reason, Novell Teaming provides multi-server Lucene configuration options that are not available in Kablink Teaming.
If you have an extremely large Novell Teaming site and you need to reindex the Teaming data, you might see improved performance by increasing these settings.
Flush threshold: The default is 100. This means that after 100 documents have been cached in memory, they are flushed to disk. Increasing the setting allows additional documents to be cached in memory before a flush occurs.
Max booleans: The default is 10000. This means that 10,000 Boolean clauses are allowed in a query. You would only need to increase this if your Teaming site includes more than 10,000 users, groups, or teams.
Max merge documents: The default is 1000. This means that the Lucene Index Server starts a new index segment after the current index segment contains 1000 documents. Only a very large Teaming site would benefit from increasing this.
Merge factor: The default is 10. This sets the number of index segments that are created on disk. When additional index segments are needed, existing segments are merged to keep the merge factor constant. Only a very large Teaming site would benefit from increasing this.
By default, the Lucene Index Server runs in the same Java Virtual Machine (JVM) as the Teaming software. To leave more memory available for the Novell Teaming software, you can configure the Lucene Index Server to run in a separate JVM. In this configuration, it communicates with the Teaming software using the RMI (Remote Method Invocation) port, which, by default is 1199.
|
ADVANCED TEAMING INSTALLATION SUMMARY SHEET |
|---|
|
Under , specify a hostname of localhost. Use the default RMI port number unless it is already in use on the server. |
To change from running the Lucene Index Server in the same JVM with Teaming to running it in its own JVM, you must complete the same steps as if you were setting up the Lucene Index Server on a remote server, including manually installing the Lucene Index Server software on the Teaming server, as described in Section 14.0, Moving the Lucene Index Server to a Remote Server.
If the Lucene Index Server requires more memory, disk space, or CPU resources than are available on the Novell Teaming server, you can move it to a remote server. This requires installing the Lucene Index Server software on that server. For instructions, see Section 14.0, Moving the Lucene Index Server to a Remote Server.
Because the availability of the index is critical to the functioning of the Novell Teaming site, you can install multiple Lucene Index Servers on multiple remote servers to provide high availability. For instructions, see Section 16.0, Running Multiple Lucene Index Servers.