SUM creates the shadow repository on the target database for all migration scenarios.
This blog provides background information as well as the current status of this new approach.
Latest News [added on June 10th 2020]
With SUM 2.0 SP 08, the approach "shadow-repository-on-target-database" is used for all migration scenarios, independent on the target release (and independent on target database type). This blog was updated accordingly
Introducing the shadow system and shadow repository
The Software Update Manager (
SUM) has several means to reduce the downtime of maintenance events like update/upgrade, database migration, or system conversion. One of these means is the
shadow system.
A shadow system basically consists of a
shadow instance and a
shadow repository:
- The shadow instance is an additional ABAP instance which is created by SUM on the application server on which the SUM was started. It is used to prepare steps executed during the downtime.
- The shadow repository exists on target product version level. This means that the shadow instance must use the target kernel to create the shadow repository with potentially new object types.
The shadow system is already created and exists during uptime processing of
SUM. As a consequence, the downtime is reduced.
[i] You can check this blog for an introduction to the shadow system, but this description does not cover the migration case.
Until SUM 2.0 SP 07,
SUM created the shadow repository on the source database specifically for the migration scenarios Database Migration Option (DMO), and System Conversion to SAP
S/4HANA for source systems with no SAP HANA database. (Note that non-migration scenarios have only one database, so that is no differentiation of source and target database is needed.)
This has changed for a system conversion to SAP S/4HANA
1909 with SUM 2.0
SP 06 (and higher):
SUM creates the shadow repository on the target database. (Again: This only applies if the source system is not yet on SAP HANA database and a database migration happens. Only then we have a source and a target database).