Loading a repository puts the objects stored in a dump file into the repository. The dump file header does not indicate the session code page in which the dump file was created. If you do not know the session code page in use when a dump file was created, do not load the dump file.
If the dump file does not include the actual content files associated with the objects you are loading, the operation reads the content from the storage areas of the dumped repository. This means that the owner of the repository that you are loading must have Read privileges at the operating system level for the storage areas in the source repository.
The load operation generates a dmi_queue_item for the dm_save event for each object of type SysObject or a subtype that is loaded into the target repository. The event is queued to the dm_fulltext_index_user user account. This ensures that the objects are added to the target repository’s index. You can turn off this behavior. For instructions, refer to Turning off save event generation during load operations.
Loading a repository is accomplished by creating and saving a load record object. The act of saving the object starts the operation.
The load operation performs periodic commits to the repository. Consequently, you cannot load a repository if you are in an explicit transaction. The Content Server does not allow you to save a load record object if you are in an explicit transaction. Similarly, you cannot perform a revert or destroy operation on a load record object if you are in an explicit transaction.