To load a dump file into a new repository:

1. Create the repository.

   Notes:

   • If the repository shares any directories with the source repository, you must assign the repository an ID that differs from the source repository ID.

   • If the old and new repositories have different owners, ensure that the new repository’s owner has Read privileges in the storage areas used by the old repository if the old repository was not dumped with the include_content property set to TRUE.

2. Create the necessary storage objects and associated location objects in your new repository.

   Each storage object in your source repository must have a storage object with the same name in the new repository. The filestore objects in the new repository must reference location objects that point to actual directories that differ from those referenced by the location objects in the source repository.

   For example, suppose you have a file store object with the name storage_1 in your source repository that points to the location object named engr_store, which references the d:\documentum\data\engr (/u04/home/installation_owner/data/engr) directory. In the new repository, you must create a file store object with the name storage_1 that references a location object that points to a different directory.

   Note

   The location objects can be named with different names or they can have the same name. Either option is acceptable.

3. If your storage areas in the source repository had associated full-text indexes, create corresponding fulltext index objects and their location objects in the new repository. Note that these have the same naming requirements as the new storage objects described in Load procedure for new repositories.

4. Create and save the following script:

   create,c,dm_load_record
   set,c,l,file_name
   full_path_of_dump file
   save,c,l
   getmessage,c

5. Log in as the owner of the installation and use IAPI to execute the script.

   When you start IAPI, connect to the new repository as a user who has Sysadmin privileges in the repository.

6. After the load completes successfully, you can destroy the load object:

   destroy,c,load_object_id

   Notes:

   • Destroying the load object cleans the load object record objects that are generated by the loading process and old state information.

   • If you created the dump file by using a script, move the dump file to archival storage or destroy it after you successfully load the file. You cannot successfully execute the script again if you leave the dump file in the location where the script created it. Content Server does not overwrite an existing dump file with another dump file of the same name.

   • If Content Server crashes during a load, you can fetch the Load Object and save it again, to restart the process. Content Server begins where it left off when the crash occurred.