62 Publishing to Oracle WebCenter Sites

This chapter provides instructions on publishing from EMC Documentum to Oracle WebCenter Sites.

This chapter contains the following sections:

62.1 Overview of the Publishing Process

Publishing from EMC Documentum to Oracle WebCenter Sites requires the CIP components Content Integration Agent and Sites Agent Services, shown in Figure 62-1, "System Architecture for Publishing to Oracle WebCenter Sites".

Figure 62-1 System Architecture for Publishing to Oracle WebCenter Sites

Description of Figure 62-1 follows
Description of ''Figure 62-1 System Architecture for Publishing to Oracle WebCenter Sites''

This section contains the following topics:

62.1.1 System Architecture and Process Flow

Content Integration Agent is used to synchronize the source and target workspaces via the cipcommander publish command, the synchronization engine, and the mappings.xml file, which provides the metadata map. Sites Agent Services exposes the web interface used by Content Integration Agent to perform the synchronization process. Following a publishing session, the synchronization process runs automatically. Details of the implementation are described below.

Initial Synchronization

When the cipcommander publish command is issued, the synchronization engine initializes the source and target by replicating metadata and associated content in this manner:

  1. The synchronization engine reads the mappings.xml file.

  2. The synchronization engine refers to the Documentum workspace (that is, cabinet or folder) named in the cipcommander publish command.

    • it reads the workspace's objects to retrieve their mapped metadata

    • it converts the metadata to WebCenter Sites-compliant format, using the mappings.xml file

  3. The synchronization engine stores the WebCenter Sites-compliant metadata to WebCenter Sites (via Sites Agent Services).

  4. The synchronization engine retrieves from Documentum the content of binary objects associated with the metadata.

  5. The synchronization engine stores the objects (via Sites Agent Services) to the target flex family.

Figure 62-2 Publishing Process

Description of Figure 62-2 follows
Description of ''Figure 62-2 Publishing Process''

Monitoring the Data Source

Following the initial synchronization, the synchronization engine starts monitoring the published cabinet or folder and automatically replicates changes to WebCenter Sites (via Sites Agent Services). For example, when an object based on the published metadata is created, modified, or deleted in the monitored cabinet or folder, the synchronization engine replicates the new or modified object, or the object's deletion to WebCenter Sites. If metadata schema is modified, the workspace must be republished.

Tuning the Integration

Content Integration Agent contains a configuration file named catalog.xml, which stores information about the publishing session. Various parameters in the file can be tuned once the synchronization process is initialized. When objects are unpublished, their information is deleted from catalog.xml . For more information about tuning, see Section 62.3.1, "Tuning the Synchronization Process."

62.1.2 Mapping Framework

The CIP mapping framework determines the success of the publishing and synchronization processes. Documentum objects can be published to WebCenter Sites as long as their metadata is mapped. The basic mapping framework for publishing involves two configurable components: the Documentum flex family and the mappings.xml file.

Documentum Flex Family

The Documentum flex family, provided with CIP, stores published objects and their object type definitions as WebCenter Sites assets. Chapter 65, "Default Mapping Specifications for Publishing" provides flex family specifications.

For simplicity, we recommend using the default flex family. Should you need to create your own flex family, refer to instructions in the Oracle Fusion Middleware WebCenter Sites Developer's Guide. Use the information in Chapter 65, "Default Mapping Specifications for Publishing" of this guide as a model of the flex family.

mappings.xml

The default mappings.xml file contains a documentum2cs section, which specifies the mappings listed below:

  • The dm_folder type maps to Documentum_Folder;dm_folder in the Documentum flex family, where

    • Documentum_Folder is a flex parent asset type that stores folder assets

    • dm_folder is a parent definition (of type Documentum Parent Definition ) that defines the folder type

  • The dm_document type maps to Documentum_Document;dm_document in the Documentum flex family, where

    • Documentum_Document is a flex asset type that stores document assets

    • dm_document is a child definition (of type Documentum Child Definition ) that defines the document type

  • Attributes are mapped in the <descriptor-mapping .../> tags. Attributes are named as listed under the Types node of the Documentum WebTop interface:

    • title

    • subject

    • keywords

    • r_version_label

    • r_full_content_size

Mapping, Publishing, and Synchronization

When publishing, bear in mind the following mapping specifications:

  • Documentum objects based on default metadata can be published to WebCenter Sites without you having to modify either the default mappings.xml file or the Documentum flex family. Running the cipcommander publish command replicates the specified workspace and its subfolders to the Documentum flex family as flex parents; documents are replicated as flex child assets. The published workspace is then monitored by the synchronization engine.

  • Publishing objects based on custom metadata (such as a new document type) requires you to first update at least the flex family with the new metadata. The mappings.xml file may or may not require updates, depending on the nature of the metadata. Details of custom mappings can be found in Section 62.2.2, "Publishing via Customized Mappings."

62.2 Publishing Procedures

This section contains the following topics:

62.2.1 Publishing via the Default Mapping

In this section, you will publish cabinets and folders to WebCenter Sites using the default mappings.xml file and Documentum flex family.

To publish via the Default Mapping

  1. Start Content Integration Agent.

    Note:

    If you changed the port in the Oracle Fusion Middleware WebCenter Sites Installation Guide, make sure that the new port is set in facilities.xml, and add -p <port> to the cipcommander publish command in 3, below (which starts CIPCommander ).

  2. Run the CIPCommander executable (located in the bin folder of the system where Content Integration Agent is installed).

  3. Publish objects of the types that are specified in the default mappings.xml file (for definitions of publishing parameters, see the table Table 62-1, "Publishing Parameters"):

    cipcommanderpublish <source_providerid> <target_providerid>-source_repname <cabinet_name>-source_path <path_in_cabinet>-target_repname <CS_content_management_site>-mapping <mapping_id>-replic_mode <full | new | updated>-bulk_resynch_interval <seconds>

    Examples:

    • To publish the Images cabinet to the CIPDemo content management site:

      cipcommander publish d7a96a63-e78c-407c-8d7f-e84988806e49 70b1e307-26a1-499c-9295-cf0b6bd01342-source_repname Images-source_path / -target_repname CIPDemo-mapping documentum2cs

    • To publish the /Sample/Trees folder in the Images cabinet to the CIPDemo content management site:

      cipcommander publish d7a96a63-e78c-407c-8d7f-e84988806e49 70b1e307-26a1-499c-9295-cf0b6bd01342-source_repname Images-source_path /Sample/Trees-target_repname CIPDemo-mapping documentum2cs

  4. When the publishing session ends, the synchronization engine starts monitoring the published object.

    Verify that modifications and deletions are replicated to WebCenter Sites (for example, modify the replicated objects, add folders and documents to the monitored object, and delete documents).

    Objects can also be unpublished. For information, see Section 62.3.2, "Unpublishing."

    To optimize the synchronization process, see Section 62.3.1, "Tuning the Synchronization Process."

    Table 62-1 Publishing Parameters

    Publishing Parameter Required Value

    <source_providerid>

    R

    Provider ID for Documentum:

    d7a96a63-e78c-407c-8d7f-e84988806e49

    <target_providerid>

    R

    Provider ID for WebCenter Sites:

    70b1e307-26a1-499c-9295-cf0b6bd01342

    -source_repname

    R

    <cabinet_name>: Name of the cabinet containing the objects to be published. Enter the name exactly as it appears in the URL.

    -source_path

    		  

    <path_in_cabinet>: Path to the object you want to publish.

    • /

    (to publish the cabinet specified by <source_repname> )

    • /<folder>/<folder>/... /<folder>

    (to publish the last folder in the path)

    -target_repname

    R

    Name of the content management site (on WebCenter Sites) on which the target flex family is enabled. Enter the site's display name, exactly as it appears on the Admin tab in the WebCenter Sites Advanced interface.

    -mapping

    R

    <mapping_id>: Value of the mapping id in mappings.xml . The value is documentum2cs .

    -replic_mode

    		  

    full | new | updated

    • full means that a full replication will be performed (by default), i.e., newly created items, updated items, and deletions.

    • new means that only newly created items will be replicated (updates and deletions will not be replicated).

    • updated means that only new and updated items will be replicated (deletions will not be replicated).

    -bulk_resynch_interval

    		  

    <seconds>: Number of seconds between two successive synchronization events.

    Default value:600

    For optimal performance, set the synchronization interval to a value that agrees with the frequency of updates to the monitored folders. For more information, see Section 62.3.1, "Tuning the Synchronization Process."

62.2.2 Publishing via Customized Mappings

If the objects you plan to publish are based on unmapped metadata, you must first map the object types.

To publish via customized mappings

  1. Depending on which type of metadata you have created, update the relevant mapping components as shown below. (We suggest reusing the Documentum flex family. The mappings.xml file is located on the Content Integration Agent host).

    New Metadata Update Guidelines
    New folder type Documentum flex family Create a parent definition for the new folder type. (The default parent definition is dm_folder .)
    New folder type mappings.xml Map the new folder type (source id ). The target id takes the value Documentum_Parent;<parent definition> (where Documentum_Parent defines the storage table for folder assets).
    New document type Documentum flex family Create a child definition for the new document type. (The default child definition is dm_document .)
    New document type mappings.xml Map the new folder type (source id ). The target id takes the value Documentum_Child;<child definition> (where Documentum_Parent defines the storage table for folder assets).
    New document attribute Documentum flex family Add the new attribute to the Documentum flex family.
    New document attribute mappings.xml Mapping the new attribute is required only if its name differs on the source and target.

    Note: Incorrect mapping of attributes does not stop the publication process, but it does produce a warning message and an entry in the log file.

  2. If you create flex filters, add the corresponding jar files to both the WebCenter Sites and the Sites Agent Services applications.

  3. Publish the objects. For instructions, see Section 62.2, "Publishing Procedures."

62.3 Maintaining the Integrated Systems

This section contains the following topics:

62.3.1 Tuning the Synchronization Process

When a cabinet or folder is published, catalog.xml is updated with data points from the cipcommander publish command. The data points identify the Documentum and WebCenter Sites systems (in the <workspace> tags) and specify replication settings (in the <replication> tag).

Following a publishing session, the synchronization engine monitors the published cabinet (folder), using catalog.xml . The BulkResynchInterval and ReplicMode parameters listed in the sample file below can be reset as shown in Table 62-1, "Publishing Parameters" (The catalog.xml file is located in the conf folder on the Content Integration Agent server).

Sample catalog.xml

<workspace id="41ce3f11-0411-46cb-b974-429162249462">
  <provider-ref refid="d7a96a63-e78c-407c-8d7f-e84988806e49" />
    <init-params>
      <param name="repname">Documents</param>
      <param name="path">/Images</param>
      <param name="repid">0c000001800045fe</param>
      <param name="itemid">0b0000018002c58e</param>
    </init-params>
</workspace>
<workspace id="6af59904-c6a3-4588-af7b-76f1422d6c10">
  <provider-ref refid="70b1e307-26a1-499c-9295-cf0b6bd01342" />
    <init-params>
      <param name="repname">FirstSiteII</param>
      <param name="repid">68ef906a-6c59-406a-84c2-b73b098cdb93</param>
    </init-params>
</workspace>
<replication>
  <link id="332e73c8-e977-4ba4-85c2-d7636281e192">
  <source-ref refid="41ce3f11-0411-46cb-b974-429162249462" />
  <target-ref refid="6af59904-c6a3-4588-af7b-76f1422d6c10" />
  <mapping-ref refid="documentum2cs" />
  <init-params>
    <param name="BulkResynchInterval">600</param>
    <param name="ReplicMode">full</param>
    <param name="IncrementalSyncDelay">10</param>
  </init-params>
  </link>
</replication>

62.3.2 Unpublishing

You can unpublish objects from catalog.xml and WebCenter Sites by executing the cipcommander unpublish command. The command clears catalog.xml of all entries that are associated with published objects (for a sample publication entry, see the sample file in "Sample catalog.xml"). The -delete parameter removes the same entries from WebCenter Sites' database.

The unpublish command takes the following form and parameters:

cipcommander unpublish <parameters>

Table 62-2 Unpublish Parameters

Unpublish Parameter Description

-all

Clears catalog.xml of all publication entries.

-linkid

Clears catalog.xml of selected publication entries.

linkid specifies the published object's link to the WebCenter Sites system. Use the value in the published object's <link> tag, which is nested within the object's <replication> tag (for sample code, see "Sample catalog.xml").For example, to unpublish an object with linkid 332e73c8-e977-4ba4-85c2-d7636281e192 from catalog.xml, run the following command:

cipcommander unpublish linkid 332e73c8-e977-4ba4-85c2-d7636281e192

-delete

Removes from WebCenter Sites' database the same objects that you are unpublishing from catalog.xml.

Legal values: <true | false>

Default value: true