Attempting to Index Long Strings of Non-Word Text Causes Laserfiche 8 Indexing to Run Slowly.

November 8, 2010 | KB: 1012704
Laserfiche 8.1.2

Summary

In order to perform a full-text search, there must be text in the repository and it must be indexed. The index process keeps track of the words contained in each document and the location of each word on the page. If a repository contains a significant number of pages that contain long strings of non-word text, indexing may run slowly. For example, if e-mails are imported into the repository that contain Base64 encoded images, Laserfiche will attempt to index long strings of text contained in the e-mails, which in many cases will cause indexing to run slowly.

Resolution

In Laserfiche 8.1.2 and higher, administrators can prevent long strings of non-word text from being indexed by completing the steps below.

Note: Laserfiche will identify any string that looks similar to Base64 encoding (and has more than 10 lines of text, as seen in the text file associated with the page in the volume) as being one that should not be indexed.

Important: The following instructions should be followed by an administrator who has experience working with Microsoft Windows' Registry Editor. In addition, since the steps involve modifying the registry, which may causes problems if not done correctly, you should create a registry backup before continuing. For more information, see Microsoft's documentation.

  1. On the machine hosting the Laserfiche Full-Text Indexing and Search Service, open the Registry Editor.
  2. Browse to the following location in the registry: HKEY_LOCAL_MACHINE\SOFTWARE\Laserfiche\LFFTS\Database\<RepositoryName>.
  3. In the left pane, right-click the registry key named after your repository, point to New, and select String Value.
  4. Name the new value CheckLongToken, then double-click it.
  5. In the Edit String dialog box, under Value data enter true and click OK.
  6. Close the Registry Editor.
  7. Open Windows' Services screen.
  8. Right-click the Laserfiche Full-Text Indexing and Search Service and select Restart.