The system automatically creates a global system temporary database (dbid 2) when the master device is created. You can create additional global temporary databases.
You can access global temporary databases from any instance in the cluster, which allows you to create temporary objects that can be shared across sessions on the same or different instances. You cannot create temporary objects such as # tables and worktables in them as global temporary databases cannot be assigned to a specific session.
Global temporary databases are identical to normal user databases with one exception: they are re-created every time the cluster starts. Global temporary databases provide standard logging, log flushing, I/O behavior, and runtime rollback features. They are supported primarily for backward compatibility with earlier versions of Adaptive Server.
All objects in the global temporary database are lost when the cluster shuts down. However, should an instance fail, the instance failover recovery step recovers the objects in global temporary databases as well, which ensures that objects created during a session on the failed instance continue to be available from surviving instances. The global temporary database ensures that temporary shareable workspace continues to exist as long as even one instance in the cluster is active.
Applications created since Adaptive Server 12.5.0.3
may create shareable tables in user-created temporary databases.
To enable existing applications to continue to work with the Cluster
Edition, user-created temporary databases that include these shareable
tables need to be dropped and recreated as global temporary databases
with the same names.
