Import/Export Manager Business Rules

All the artifact numbers (change request number, requirement number, etc) are remapped to a new number in the target. If the change request has been shared across multiple views or sand boxes, the change request number will remain intact. The older change request number will be stored in a custom field for cross referencing purposes.

The export process evaluates all the shares and links that cross the selected project boundaries. Shares and links that cross project boundaries, but are within the list of exported projects, are not broken by Import/Export Manager.
For those shares whose parent or root is in a different project which is not being exported, Import/Export Manager will make this item its own root. It will reset the branching behavior to model a root item.
Links with either the source endpoint or the target endpoint outside the project scope will be ignored and will not be exported.

Cross project shares are shares whose parents reside outside the selected projects boundaries. These shares and links are broken by Import/Export Manager and reset to be root items. The branching behavior of these items is reset to model a root item. If the source item is pinned, the pinned behavior is carried over to the target.

Custom component names are matched by their internal name during import. If there is a match found on the target server for a given component name, the source and target components are considered to be the same. If a match is not found, a new custom component will be created on the target server and all its associated data, properties, and enumerations will be copied over. It is the responsibility of the user to review their source and target components and see if there are matching custom components which are not related but happen to have the same internal name. In this case, the user will have to rename either the source/target component before import.

For components on the source and target server that match, the import operation iterates through all the custom properties available and checks to see if a custom property with the same internal name for the component exists on the target server. If it is not found, this property will be created on the target server.

Microsoft SQL Server has limitations with the row length and importing a large number of custom properties could potentially cross the row length limit. One way to avoid this issue is to exclude disabled properties from being carried over. The Exclude disabled properties check box is provided during export. This will ensure that all the catalog field/enumeration entries and the values for the disabled properties on the source server are not imported.

Before custom properties are added to the target server, Import/Export Manager will do a row length check to determine if the target database table allows new columns to be added. Import/Export Manager will not continue if this check fails.

Please also consider:

Property with the same internal name but different enumerations: If two enumerated (single/multi-select) custom fields have the same internal name and the same type, the target server configuration must have the enumerations found on the source as well as those found on the target. For example, if the source field has 100 red, 200 blue, and 300 white and the target field has 100 blue and 200 orange, Import/Export Manager will add red and white to the list of enumerations for the target field. Import/Export Manager will see that the 200 blue is equivalent to the 100 blue and ignore orange, but it needs to find red and white, regardless of their code numbers. After the import, this field will have the following enumerations on target server configuration for this example field: 100 blue, 200 orange, <some number> red, <some number> white. Import/Export Manager preserves the values (visible to users) and does not worry about the code numbers. However, you may want to set up the extra enumerations needed on the target server manually before importing to control those codes. Disabled enumerations are still copied for enabled custom fields because some items may have those values. The enumerations will be disabled on the target unless the target field already has an equivalent enabled enumeration. The default values for an enumerated property are not merged.
Custom property being added to the target already exists with the same data type but different length: The target custom property's length will be increased to prevent data truncation issues for the imported data.
Custom property being added to the target already exists with a different data type: Import/Export Manager will create a new custom property with a system generated name.

Import/Export Manager does not delete the Projects selected for export from the source server. It is up to the user to decide based on their requirements.

Filters, queries, and audits will not be imported by default. This can be changed in the user interface.

In addition to the database metadata, the Import/Export Manager also copies over vault contents to the target server. If the target server has multiple hives, the contents are added to all the hives in a round-robin fashion, similar to the way StarTeam Server processes content. The Import/Export Manager does an up front estimation of the total size of the vault content being exported from the source and checks to see if the target hive has the capacity to handle importing the contents. If there is not sufficient storage space, the Import/Export Manager will throw an error message and abort the operation before importing any data. Estimation of space on the target hives depends on various parameters like storage threshold limit of the hive, disk partition capacity, and also the Allow new Archives option. It is recommended that you ensure that there is ample disk space on the target machine and that the hives are configured appropriately before starting import.

File contents, attachments, and contents from content properties are copied over to the target as is, without any processing. Change package attachments, however, need to be processed before being imported to the target server. Change package attachments have ids embedded in the content. They need to be re-mapped to the corresponding id in the target. The imported contents will therefore end up having a different md5 value compared to the one in source. This is essential so that the imported change packages can be opened and used.

Import/Export Manager does not copy over the source server's licensing information. It is important for the user to ensure that the target server is licensed appropriately. For example, if the source server has enterprise advantage license, and the user is importing requirements data from this server to the target server, it is required that the target server has an enterprise advantage license to view the imported requirements.

Source and Target StarTeam Servers must be running the same build number.

If a tip item is moved across project boundaries, the historical revisions of the item will be available in the target to support rolled back configurations. The tip item will not be available in the target since it does not belong to the imported project.