Abstract: A recovery of a secondary DB is executed by a log including a update difference of a primary DB, and a command is executed by adding the command to the log and analyzing the log by a secondary site. An operation command execution in the secondary site is applied to the DB having the consistency in the transaction at the same timing as a primary site or an intended timing, by transferring and executing the operation command via the log. In the case that the command is constituted by a snap shot generating instruction, a plurality of volumes of a mirror set forming a secondary storage apparatus are set to a pair state and the secondary DB is written in each of the volumes and synchronized, and the mirror set is set to a split state after the synchronization is finished, whereby a secondary DB is stored.

Abstract: A technique capable of constructing a disaster recovery system reduced in performance degradation of a primary system is provided. The technique includes a step of conducting synchronous writing of log information into a secondary storage subsystem in a secondary system when a write request received from a host computer is a write request of log information, a step of temporarily storing a write request and conducting asynchronous writing into the secondary storage subsystem when the received write request is a write request of database data or status information, a step of modifying log information, data in a database area, and status information in the secondary storage subsystem according to contents of a write request received from a primary storage subsystem, and a step of recovering the database area according to contents of log information in a location indicated by the status information.

Abstract: When a database area multi-written or replicated in an external storage unit is separated, a database management system writes consistent data into the external storage unit. A method and system capable of controlling a timing of separating the replicated database area is provided. Furthermore, new data directory information necessary to access the replicated database is dynamically added based on the data directory information of the replication source database. A method and system capable of selecting whether to access the database which is the source of the replication or access the replicated database are provided. Thus, when the database area multi-written or replicated in an external storage unit is separated and physical replicated databases are created, it is possible to maintain consistency of the replicated databases while continuing transaction processing.

Abstract: To reduce operational and management costs during normal operations while recovering a database without loss and maintaining on-line performance on a site. A first system includes a primary storage system (103) that stores a DB (107) and a main computer (100) that executes a primary DBMS (101), which provides a DB. A second system includes a secondary (113) that receives from the primary storage system (103) a copy of a log, which shows update differences of the DB (107), and stores a secondary DBMS (117), and a subset (500) that recovers the secondary DB (117) according to the log that is copied from the primary storage system (103). When a failure occurs in the first system, the first system is switched to the second system. A second computer (110) that executes a second DBMS (111) is added to the second system, and the secondary DB (117) that is recovered or is being recovered in the subset (500) is taken over to the second computer (110).

Abstract: If relocation of a table with a high reference frequency is made possible, the database management method and system are arranged to improve performance according to the number of increased nodes and accumulate a reference frequency of each external storage unit allocated to each database management node. A data storage area for storing the table with a high reference frequency, located in the external storage unit, is specified as a shared data area to be referenced by plural nodes. The system is controlled to cause the nodes to refer to the shared table stored in the shared data area on any timing in parallel. This prevents load from being concentrated onto the node allocated to the data area.

Abstract: A record having a fixed length data portion and a variable length data portion is stored in memory units, each capable of being inputted/outputted independently, in unit of sub-record while the fixed length data portion is divided into a plurality of fixed length sub-records each having the same length and the variable length data portion is made one variable length sub-record. When data operation of the record such as retrieval, update and deletion is required, sub-records necessary for the data operation are limited and the data operation is made in a unit of one or more sub-records, so that data processing of unnecessary sub-records is not made.

Abstract: A disaster recovery system and a method therefor, having: a function for receiving write requests from a host computer to write log information indicating the contents of a database process executed for a database buffer on a main site, database data updated on the database buffer and status information indicating the position of log information used for disaster recovery; a function for transferring the received write request for log information to a recovery site by a synchronous remote copy process; and a function for transferring the received write request for database data to the recovery site by an asynchronous remote copy process. While remote copy can be executed at a long distance without data loss, the deterioration of the performance of a database process on a main site can be suppressed.

Abstract: There is provided a technique capable of executing a predetermined processing while continuing a database processing. Firstly, the database to be operated is switched from the main database to a duplicate database. The predetermined processing is executed to the main database which becomes the non-operation mode by the switching. Next, a database access request is distributed to the duplicate database which becomes the operation mode by the switching, thereby performing the database processing. Update information indicating the content of the update processing performed to the duplicate database is output. The output update information is reflected on the main data base for which the predetermined processing has been executed. After this, the database to be operated is switched from the duplicate database to the main database.

Abstract: A technique capable of constructing a disaster recovery system reduced in performance degradation of a primary system is provided. The technique includes a step of conducting synchronous writing of log information into a secondary storage subsystem in a secondary system when a write request received from a host computer is a write request of log information, a step of temporarily storing a write request and conducting asynchronous writing into the secondary storage subsystem when the received write request is a write request of database data or status information, a step of modifying log information, data in a database area, and status information in the secondary storage subsystem according to contents of a write request received from a primary storage subsystem, and a step of recovering the database area according to contents of log information in a location indicated by the status information.

Abstract: A parallel database management system with a database divided into a plurality of external memory units is disclosed. A control CPU, upon receipt of a request for access to the database, decides whether the access request is for storing or searching a record. For a record storage request, the control CPU determines an external memory unit to store the record using a hash function. In the case of a record search request, on the other hand, all the external memory units are searched in parallel without using the hash function thereby to obtain the desired record. The hash function is altered upon addition of an external memory unit.

Abstract: A parallel database management system with a database divided into a plurality of external memory units is disclosed. A control CPU, upon receipt of a request for access to the database, decides whether the access request is for storing or searching a record. For a record storage request, the control CPU determines an external memory unit to store the record using a hash function. In the case of a record search request, on the other hand, all the external memory units are searched in parallel without using the hash function thereby to obtain the desired record. The hash function is altered upon addition of an external memory unit.

Abstract: A parallel data base management system with a data base divided into a plurality of external memory units is disclosed. A control CPU, upon receipt of a request for access to the data base, decides whether the access request is for storing or searching a record. For a record storage request, the control CPU determines an external memory unit to store the record using a hash function. In the case of a record search request, on the other hand, all the external memory units are searched in parallel without using the hash function thereby to obtain the desired record. The hash function is altered upon addition of an external memory unit.

Abstract: When a database area multi-written or replicated in an external storage unit is separated, a database management system writes consistent data into the external storage unit. A method and system capable of controlling a timing of separating the replicated database area is provided. Furthermore, new data directory information necessary to access the replicated database is dynamically added based on the data directory information of the replication source database. A method and system capable of selecting whether to access the database which is the source of the replication or access the replicated database are provided. Thus, when the database area multi-written or replicated in an external storage unit is separated and physical replicated databases are created, it is possible to maintain consistency of the replicated databases while continuing transaction processing.

Abstract: A record having a fixed length data portion and a variable length data portion is stored in memory units, each capable of being inputted/outputted independently, in unit of sub-record while the fixed length data portion is divided into a plurality of fixed length sub-records each having the same length and the variable length data portion is made one variable length sub-record. When data operation of the record such as retrieval, update and deletion is required, sub-records necessary for the data operation are limited and the data operation is made in a unit of one or more sub-records, so that data processing of unnecessary sub-records is not made.

Abstract: A record having a fixed length data portion and a variable length data portion is stored in memory units, each capable of being inputted/outputted independently, in unit of sub-record while the fixed length data portion is divided into a plurality of fixed length sub-records each having the same length and the variable length data portion is made one variable length sub-record. When data operation of the record such as retrieval, update and deletion is required, sub-records necessary for the data operation are limited and the data operation is made in a unit of one or more sub-records, so that data processing of unnecessary sub-records is not made.

Abstract: A parallel data base management system with a data base divided into a plurality of external memory units is disclosed. A control CPU, upon receipt of a request for access to the data base, decides whether the access request is for storing or searching a record. For a record storage request, the control CPU determines an external memory unit to store the record using a hash function. In the case of a record search request, on the other hand, all the external memory units are searched in parallel without using the hash function thereby to obtain the desired record. The hash function is altered upon addition of an external memory unit.

Abstract: A parallel data base management system with a data base divided into a plurality of external memory units is disclosed. A control CPU, upon receipt of a request for access to the data base, decides whether the access request is for storing or searching a record. For a record storage request, the control CPU determines an external memory unit to store the record using a hash function. In the case of a record search request, on the other hand, all the external memory units are searched in parallel without using the hash function thereby to obtain the desired record. The hash function is altered upon addition of an external memory unit.

Abstract: A record having a fixed length data portion and a variable length data portion is stored in memory units, each capable of being inputted/outputted independently, in unit of sub-record while the fixed length data portion is divided into a plurality of fixed length sub-records each having the same length and the variable length data portion is made one variable length sub-record. When data operation of the record such as retrieval, update and deletion is required, sub-records necessary for the data operation are limited and the data operation is made in a unit of one or more sub-records, so that data processing of unnecessary sub-records is not made.

Abstract: A dynamic file expansion method and system for on-line database systems is provided. When an empty space of a file becomes insufficient, the file is dynamically expanded asynchronously with a record input/output process relative to an already existing space, by using another process while continuing the record input/output process. Jobs are not therefore intercepted and adversely affected by insufficient empty space of the file. In the structure of the method and system, a file expansion timing detecting section monitors the remaining empty space of a file with the help of a record input/output processing section. When the remaining empty space becomes equal to or less than a predetermined space registered in a file expansion information registering section, a dynamic file expansion is requested to an asynchronous file expanding section.

Abstract: A record having a fixed length data portion and a variable length data portion is stored in memory units, each capable of being inputted/outputted independently, in unit of sub-record while the fixed length data portion is divided into a plurality of fixed length sub-records each having the same length and the variable length data portion is made one variable length sub-record. When data operation of the record such as retrieval, update and deletion is required, sub-records necessary for the data operation are limited and the data operation is made in a unit of one or more sub-records, so that data processing of unnecessary sub-records is not made.