Oracle International Corporation

Philip Yam

Kumar Rajamani

Jaebock Lee

Andre Kruglikov

Thomas Baby

Chandrasekharan Iyer

Hochak Hung

William H Waddington

Yunrui Li

Shrikanth Shankar

Giridhar Ravipati

Sanket Jain

David M Alpern

Guhan Viswanathan

Alan Choi

Techniques are provided for diverting, to cloned metadata, live access to original metadata of an application container that is being concurrently upgraded. In an embodiment, a database server stores, within an application container of an application, original metadata that defines objects for use by pluggable databases of the application. The database server receives a maintenance request to adjust the original metadata. The database server creates, in response to receiving the maintenance request, a reference container that contains cloned metadata that is a copy of the original metadata. The database server receives, during or after creating the reference container, a read request to read one of the objects. The database server concurrently performs both of: executing the maintenance request upon the original metadata, and executing the read request upon the cloned metadata of the reference container.

Asynchronous shared application upgrade

A database server instance automatically detects configuration issues when a pluggable database (PDB) is plugged into a new destination container database (CDB). The database server instance identifies one or more update templates, within the destination CDB, that, when run over the relocated PDB, will update the configuration of the PDB to conform to the configuration of the destination CDB. Instead of requiring an administrator to initiate update scripts from a DBMS kernel to reconfigure a PDB, the DBMS creates the update templates by recording commands run within PDBs in connection with system updates. These recorded update templates may then be run over relocated PDBs, to configure the PDBs according to the configuration of the destination CDB. Further, the update templates may be pre-recorded update templates, which record commands to perform configuration updates, to PDBs, that have never before been performed within the CDB.

Automatic reconfiguration of relocated pluggable databases

In an approach, memory usage within a database management system is optimized by designating an application root which stores data and metadata for database objects that are common across pluggable databases which support the same application. In addition, the application root can be used to manage access to the pluggable databases by defining common users which are created at the application root, but can be used to access any of the member pluggable databases. Furthermore, different versions of the same application can be supported concurrently by maintaining separate application roots for each version. For instance, when an upgrade is applied to the application root, a clone can be created and the update can be applied to the clone. Then, the member pluggable databases which are to be upgraded can be synchronized to the clone, while leaving other pluggable databases servicing the previous version of the application.

Application containers in container databases

A method and apparatus for efficiently managing cached objects is provided. A mapping is a typed relationship between cached objects. A base object contains content, and a mapped object is mapped to a base object when there is a particular relationship between the mapped object and the base object. The type of mapping defines how the objects in the relationship are treated when the mapped object is created, read, or written. At creation time, the mapping type may define how the mapping relationship is recognized. Sometimes recognizing a relationship requires establishing equivalence between the objects. At read and write time, the mapping type may define on which object a mutual exclusion lock is held, which content is returned or which object's content is updated. A wide range of applications can benefit from enhanced object caching performance, and in addition, application may define application-specific semantics on mapping types as well.

Memory-mapped objects

A container database may contain multiple database dictionaries, each database dictionary defining a pluggable database. When database sessions are established on a container DBMS, each database session is given access to a pluggable database by establishing the respective database dictionary of the pluggable database as the database dictionary for that database session. Database commands issued through database session can only access the database objects defined in the database dictionary established for the database session.

Container database

A database may facilitate zero-downtime upgrades by concurrently maintaining multiple editions of database objects for use by both pre-upgrade and post-upgrade clients of a database application. Operations performed within the database are associated with an edition based on, for example, an initiating client or transaction. When an operation references an object or data, the database automatically performs the operation using the object or data associated with the edition with which the operation is itself associated. The database may determine the associated edition without explicit identification of the associated edition in a query or in code. Thus, no client or stored procedure code changes are necessary to reflect a new edition added during an update. Data changes in one edition may be automatically and immediately propagated to the other edition through the use of cross-edition triggers, thereby allowing both pre-upgrade and post-upgrade clients to remain fully functional throughout an upgrade.

Low-downtime and zero-downtime upgrades of database-centric applications

Disclosed are improved methods, systems, and media for cursor sharing and cursor pruning. According to some approaches, existing child cursors are distinguished using a cursor sharing criteria node structure. A sharing criteria node (also referred to as a 'diagnostic' node) is created when a sharing criteria mismatch/failure happens. The node contains information about why the child cursor could not be shared and also information that can be used to quickly re-evaluate this sharing criterion in the future.

Growing community of inventors

San Carlos, CA, United States of America

Philip Yam