Windows Server 2003 : Using the Indexing Service - Creating and Configuring Catalogs

A catalog contains all the index information for a particular set of file directories. During installation, the Indexing Service creates a default catalog called System. This catalog lists the contents of all permanently attached disk drives and, by default, all the directories and subdirectories on the drives. If IIS is installed, the Indexing Service also creates a Web catalog that contains all the IIS files.

You can create catalogs, adding and removing them as needed. You can also configure catalogs, setting which directories are to be included or excluded and specifying which properties are to be stored.

Creating a Catalog

To create a catalog for the Indexing Service, open the MMC with the Indexing Service snap-in and follow these steps:

1.	Highlight Indexing Service in the console tree.
2.	From the Action menu, choose New, and then choose Catalog.
3.	In the Add Catalog dialog box, supply a name for the catalog and a path to the folder in which you want the catalog placed. (See Figure 1.) Click OK. Figure 1. Creating a new catalog
4.	You must stop and restart the Indexing Service before the new catalog can be found and indexed. Right-click Indexing Service, and choose Stop from the shortcut menu. To restart the Indexing Service, right-click Indexing Service again and choose Start from the shortcut menu.

Note

Catalogs can’t be added to a remote computer if the default administrative shares have been removed.

Configuring a Catalog

After you create a catalog, you need to configure it so that it works as you expect. To do so, open the Indexing Service console and locate the catalog. Right-click the catalog, and choose Properties to do the following:

• Index a Web server Click the Tracking tab and, in the WWW Server box, select the Web server you want to index. If IIS isn’t installed, this option isn’t available.

• Index files with unknown extensions Click the Generation tab. Ordinarily, this setting is inherited from the overall Indexing Service properties and is inactive by default. If you want all the files in this catalog to be indexed, including those without installed filters, clear the Inherit Above Settings From Service check box. Select the Index Files With Unknown Extensions check box.

• Generate abstracts Click the Generation tab. The Generate Abstracts setting is inherited from the overall Indexing Service properties and is inactive by default. To select this option, you must first clear the Inherit Settings From Service check box. If Generate Abstracts is selected, the Indexing Service produces abstracts in the list of query results. This slows the query process, so it’s best not to increase the default size of abstracts.

• Add a network share alias automatically Click the Tracking tab. By default, this setting is inherited from the Indexing Service, where it is enabled.

With the exception of Generate Abstracts, all these changes take effect only after you stop and restart the catalog. To do so, right-click the catalog, point to All Tasks on the shortcut menu, and choose Stop. Then right-click again, point to All Tasks, and choose Start. If you change the Generate Abstracts setting, you need to stop and restart the Indexing Service for the change to be recognized.

Including or Excluding a Directory

By default, the System catalog includes everything on the local drives, excluding temporary Internet files and history files. When you create a new catalog, you have to add the directories that are to be included as well as specifically exclude directories that are not to be part of the index. To add a directory to a catalog, follow these steps:

1.	Open the Indexing Service console. Right-click the new catalog, and select new and then Directory from the shortcut menu.
2.	In the Add Directory dialog box, supply the path to the directory and the Uniform Naming Convention (UNC) path, if necessary.
3.	If the directory is on another computer, supply a name and password for a user with permission to access the remote share, as shown in Figure 2. Click OK and the directory becomes part of the catalog. Figure 2. Adding a directory on a remote computer to the catalog

To change settings for a directory, double-click the directory in the details pane of the Indexing Service console to open the Add Directory dialog box.

To exclude a particular directory, you must specify it. For example, the Bettiman Archives catalog shown being created in Figure 31-4 includes a directory called Book Files. In that directory is a subdirectory called Correspondence that you want to exclude from indexing.

To exclude a directory, follow these steps:

1.	Open the Indexing Service console. Right-click the catalog, and select New and then Directory.
2.	In the Add Directory dialog box, supply the path to the directory you want to exclude and the UNC path, if necessary. Again, if the directory is on another computer, supply a name and password for a user with permission to access the remote share.
3.	In the Include In Index area, select No. Click OK.

The directory appears in the directory list, but under Include In Catalog, the entry is No.

Note

Although you can include a directory and then specifically exclude a portion of it, this process does not work in reverse. If you exclude a directory, you cannot then include some portion of it, even if you specify the directory and set it to be included in the index. If you attempt to do this, the directory is in the catalog’s directories and is listed as being included in the catalog, but it is not indexed.

Indexing and Security The Indexing Service is fully compatible with NTFS security. If the catalog is on an NTFS volume, users don’t see documents in the results list unless they have permission to do so. However, if you index a UNC share, the results list shows the documents on that share whether or not the user has permission to access the documents. Users are also able to see a catalog on a FAT drive whether they have permissions or not. If the remote FAT volume isn’t hosted by Microsoft Windows NT, Windows 2000, or Windows Server 2003, the system is forced to scan the volume periodically for changes. Remote Novell NetWare and UNIX shares can be indexed; however, there is no security checking. Novell NetWare volumes must be periodically rescanned to detect changes. Encrypted documents are not indexed. If a document in the index is later encrypted, it is removed from the index.

Configuring the Property Cache

The Indexing Service saves certain file properties in a two-level cache in each catalog. The primary level contains a small number of values that are accessed frequently. The secondary level contains values that are used less often. Table 1 shows the properties that are stored in each catalog by default, though many more are available.

Table 1. Property values stored in a catalog by default

Friendly Name	Function	Value	Storage Level
DocTitle	Document title	0x2	Secondary
	Unique identifier for NTFS volumes	0x5	Primary
	Work ID of the parent directory	0x6	Primary
	Secondary storage ID, used internally by the Indexing Service	0x7	Primary
File Index	Unique identifier of a document in an NTFS partition	0x8	Primary
Path	Document path	0xb	Secondary
Size	Document size	0xc	Secondary
Attrib	Document attributes	0xd	Primary
Write	Date and time the document was last written to	0xe	Secondary

In general, approach changing these properties with caution, always bearing in mind the following facts:

• Adding property values to either level—but particularly to the primary level—has a negative effect on the performance of the Indexing Service.

• Adding variable-length properties to the primary level increases the size of the cache exponentially.

• After you add a property value to either level and then restart indexing, you can’t change the level for that property.

But don’t assume that changing the property cache is always a poor idea. For example, information such as when files were created or when they were last accessed can be important, depending on your needs.

Adding a Property

To add a property to the properties saved in the property cache, follow these steps:

1.	Open the Indexing Service console. Under the appropriate catalog, click Properties.
2.	In the details pane, select the property you want to add.
3.	From the Action menu, choose Properties to open the property’s Properties dialog box.
4.	To include this property in the property cache, select the Cached check box. (See Figure 3.) You can see and change the datatype and the size of the property. (Only properties with variable sizes can be adjusted.) The storage level can also be assigned. Click OK when you’re done. Figure 3. Adding a property to be cached

Changes take effect after the Indexing Service is stopped and restarted, but these newly added properties are included in the property cache only for new documents. To update the entire index with the newly included properties, perform a full scan of the index, as described in the next section. If you later decide to remove a property or alter its settings, you can do so by clearing the Cached check box in the property’s Properties dialog box. Again, a full scan is required to update the entire index.

Running a Scan of the Index

A full scan of the index consists of a complete inventory of all the documents in the catalog. The Indexing Service automatically performs a full scan when it is first installed, when a directory is added to a catalog, and as a part of recovery if an error occurs. Incremental scans are done automatically when the Indexing Service restarts to detect documents that were changed while it was inactive. You can perform a full or incremental scan at any time by following these steps:

1.	Open the Indexing Service console. In the console tree, click the appropriate catalog and then click Directories.
2.	In the details pane, select the directory to be scanned.
3.	From the Action menu, point to All Tasks and then choose Rescan (Full) or Rescan (Incremental), depending on the type of scan you want to perform. You’re asked to confirm your choice. Click Yes and the scan proceeds.