Maximize the Performance of CFS
The performance of CFS varies greatly depending on the import and index task A processing task run by CFS on documents. Update-index tasks run when a document's metadata (but not its content) is updated. Delete-index tasks run when a document is deleted from a repository.s that you run. Running large numbers of tasks can reduce performance, and tasks such as Eduction consume significant resources.
This section describes how you can improve the performance of CFS.
Install CFS on a Separate Machine or Disk
Depending on the system resources used by your connector, you might be able to run CFS using the same machine and hard disk. However, for the best performance, install CFS on a separate machine or disk to the connector.
Optimize the Number of Processing Threads
You can configure the number of processing threads used by CFS by setting the ThreadCount
configuration parameter. The optimum setting is double the number of free processor cores.
If you have a quad core machine, set the number of threads to 8.
Send Documents to IDOL in Large Batches using a Shared Folder
CFS sends processed documents to IDOL in batches. Indexing is more efficient with larger batches. You can set the batch size using the configuration parameter IndexBatchSize
. The default batch size is 100 documents.
By default, CFS sends index data to IDOL using sockets. If a shared disk is available, consider setting the configuration parameter IndexOverSocket
to False
. When this parameter is set to False
, CFS sends only filenames to IDOL, rather than all the index data. IDOL reads the data from the shared folder.
Configure Logging
The amount of logging that occurs can have a significant impact on performance. In normal operation, set the LogLevel
parameter to Normal
. Setting LogLevel
to Full
reduces the performance.