International Business Machines Corporation

Larry William Youngren

Gary Ross Ricard

Adam Thomas Stallman

John J Vriezen

David Finian Owen

Chad Allen Olstad

David Lee Hermsmeier

Wilson D Lee

Peter B Passe

Mark John Anderson

Richard Lee Cole

Michael Joseph Corrigan

Ruth Janine Poole

Dennis Steven DeLorme

Michael Lawrence Nordstrom

David Rolland Welsh

George D Timms, Jr

Richard Miles Rocheleau

William S Davidson

Mark L Holm

Margaret Ellen Levering

A minimized journaling mechanism stores minimized journal data in a format that allows for display and outputting the journal data in human-readable form. When a change to a record occurs, instead of writing only the changed bytes, all of the bytes in each field that changed are written to the journal, along with all of the bytes in each field selected to be always journaled. A default object is created with default data in all of the fields. When the journal entry needs to be output in human-readable form, the default object is read, and the minimized journal entry is then overlaid on the default object. The result is an object that contains default data in all non-selected fields that were not changed, with the journal data in all fields and that did change and in all fields that were selected to always be journaled.

Journaling database changes using minimized journal entries that may be output in human-readable form

The disclosure and claims herein are directed to efficient journaling for recovery of a database index by journaling zones of a page. A journal mechanism maintains a page zone bit map that includes a bit for a plurality of zones in each page to indicate which zones have had their unchanged image journaled before being changed since a last sync point update. The page zone bit map has a bit for each zone in each page so that the status of each zone can be tracked separately. Tracking the smaller zones of the pages makes the process more efficient both at run time and during recovery by reducing the period of time for memory deposits and reducing the amount of total redundant/recovery data sent to disk for larger pages.

Journaling database changes using a bit map for zones defined in each page

A method, apparatus, system, and signal-bearing medium that in an embodiment determine, after a restore of a version of an object, where to begin applying changes from a change log to the object based on an identification in the saved version of the object. When a save command for an object is received, an identification of the change log entry associated with the save command is stored in the saved version of the object. Changes to the object then continue to be logged to the change log. After the saved version of the object is restored, the identification in the saved version of the object is used to find the change log entry at which to start applying changes from the change log to the object. In this way, the point in the change log at which to start applying changes to the object may be determined based on the saved version of the object. In an embodiment, this allows the object to continue to be accessed while the save operation is performed. Further, this prevents confusion as to which starting point in the change log to use when multiple save operations have been performed.

Storing object recovery information within the object

An apparatus and method includes a database journal that records journal entries. Each journal entry includes a validation value that is computed from the data and metadata for the record before it is changed. In this manner, journal entries may be applied with confidence to a record in a database. This is especially useful when minimized data journal entries are used. If the validation value of the journal entry matches the validation value computed from the data and metadata of the record that the journal entry is to be applied to, then the record before applying the journal entry is identical to the record in the original database, and the journal data may be applied to the database record. If the validation value of the journal entry does not match the validation value computed from the data and metadata of the record that the journal entry is to be applied to, the record is not the same as in the original database when the journal entry was generated. As a result, an error message is generated to indicate that an error was encountered when applying the journal entry. In this manner, the apparatus and method of the preferred embodiments automatically validate a database record before applying journal data to the record.

Apparatus and method for validating a database record before applying journal data

A database manager selectively caches journal entries for certain database changes based on dynamic selection criteria. Preferably, a journal disk drive is provided for storing changes to the database. Database change entries are sequentially cached in a main memory journal buffer, awaiting journal write. When a buffer is filled, its buffer contents are written to the journal disk. However, certain change operations will cause the contents of the buffer to be written to the journal disk immediately, before filling the buffer. Certain database change entries selectively cause the buffer to be written, depending on dynamic selection criteria. Preferably, these selection criteria take into account both the estimated time to recover in the event of a system crash and the current level of activity on the system. Generally, the busier the system, the more data will be cached, and vice versa.

Method and apparatus for selective caching of transactions in a computer system

An apparatus, program product and method manage a queue through the use of element-based journaling to record changes made to logical elements in a queue. In contrast to convention memory-based journaling, where any changes to the memory representing an element in a queue are journaled, only those operations that affect the logical ordering and/or placement of an element on a queue, or the logical contents of such an element, are journaled. Memory management operations such as pointer manipulation operations that modify pointers or other memory management data, but do not otherwise modify the actual elements in a queue or their relative ordering, need not be journaled. As a consequence, the storage and processing overhead associated with journaling may be substantially reduced, thereby substantially reducing the overhead associated with maintenance and recovery of a queue.

Automated queue recovery using element- based journaling

System control over the logging of objects in order to meet the user specified recovery requirements. Under the fixed recovery time environment, the user chooses a length of time to be spent on object recovery, and the system dynamically manages the logging of objects to meet this time. The shorter the time the user chooses, the more objects the system will log, and the more performance degradation there will be as a result of the logging at run-time. The user may partition storage into Auxiliary Storage Pools (ASPs), which are groups of non-volatile storage, and then specify the recovery time on a per ASP basis. Under the minimal impact environment, the system dynamically manages the objects to be logged such that the object logging has a minimal impact on run-time performance.

System managed logging of objects to speed recovery processing

An in use table manager in a computer system uses an in use table to track the use of files, or objects. The in use table is used to determine which objects may need recovery in the event of a system failure. Object addresses are hashed by the in use manager to identify a preferred slot in the table. The slots contain information identifying the object, and indicating the extent of use of the object. The in use manager assigns alternate slots, and dynamically changes the size of the in use table to reduce contention for slots. Several atomic operations on the table ensure integrity of the table, while permitting concurrent use. Portions of the table are bundled into single I/O operations to enhance system performance by minimizing I/O.

High concurrency in use manager

Optimization I/O buffers are used to write data into or read data from auxiliary storage using a double buffering scheme. When a user wants to send data to auxiliary storage, the first buffer is filled with data and the contents are asynchronously written to auxiliary storage. Without waiting for the asynchronous write to complete, the second buffer is filled with data. When the second buffer is filled, the contents of the second buffer are asynchronously written to auxiliary storage. If the first buffer has not completed its write operation by this point, the buffers are determined to be too small and the size of both buffers is increased. Control switches back and forth between these two buffers until all of the desired data is sent to auxiliary storage. The size of the buffers is increased until the computer system does not have to synchronously wait for one buffer to complete its write operation before it can refill that buffer with new data. Ideally, the first buffer should complete its write operation when the second buffer is filled beyond a predetermined threshold percentage, such as one half. If the first buffer completes its write operation before the second buffer is filled beyond the predetermined threshold percentage, the buffers are determined to be too large and the size of both buffers are decreased. The size of the buffers is dynamically increased and decreased as needed to achieve an optimum size based on the unique and dynamically changing computer environment.

Optimized I/O buffers having the ability to increase or decrease in size

A quick recovery of logical files which provide alternate views of databases is provided. Unchanged logical file pages are journaled before being changed. Transactions affecting databases covered by the logical files are also journaled. To recover a logical file, the journaled unchanged pages of the logical file that correspond to the changed pages are inserted back into the logical file, and the transactions that were journaled are processed to provide the changes to the logical file and to the database. This brings the logical file and the underlying database up to date, and in synchronization with each other.

Database index journaling for enhanced recovery

A key estimator estimates the number of keys over a key range defined by key endpoints in an index to a data space. The number of keys in the key range required to be processed for a particular operation is estimated as a function of the number of pages referenced during a range level limited search. Two keys defining range endpoint keys are searched down to their lowest level in the tree. The level limit is then calculated as a function of desired granularity or accuracy of the estimate. The entire range of keys in the desired key range is then searched down to the level limit and the number of pages referenced during the search is counted and multiplied by an average key density per page to calculate the number of keys in the range.

Growing community of inventors

Rochester, MN, United States of America

Larry William Youngren