Database administration for data sharing

Using data sharing has implications for planning exit routines, authorizing users, and loading and reorganizing data.

Because the DB2® catalog is shared by all members of a data sharing group, data definition, authorization, and control are the same as for non–data sharing environments. Be sure that every object has a unique name, and be sure that the shared data resides on shared disks.

Planning for exit routines

If you use exit routines, such as a field or validation procedure or the access control authorization routine, ensure that all members of the group use the same routines.
Recommendation: Place all exit routines in a program library that is shared by all members of the group.

Authorizing users

Use the same authorization mechanisms that are in place for non–data sharing DB2 subsystems to control access to shared DB2 data and to members. Because all members in the group share the same DB2 catalog, an authorization ID has the same granted privileges and authorities for every member of the group.

As suggested for non–data sharing DB2 subsystems, use a security system outside of DB2 (such as RACF®® or its equivalent) to control which user IDs can access which members. RACF, for example, does not recognize a data sharing group as a single resource. Therefore, you must separately define DB2 resources to RACF for each member of the group, and connect all user IDs to a RACF group that permits access to all those resources. Or you can permit separate groups of user IDs to access different sets of resources. (In the latter case, however, you cannot move work freely among all members of the data sharing group.)

Each member of a data sharing group uses the same names for the connection and sign-on exit routines. As a good practice, all members of a group should share the same exit routines. Sharing avoids authorization anomalies such as:
  • Primary authorization IDs that are treated differently by different members of the group
  • Primary authorization IDs that are associated with different sets of secondary IDs by different members of the group

Loading and reorganizing data

You can load or reorganize data from any member of a data sharing group.