High availability database management system and database management method using same

摘要

Provided is a method of managing a database (DB) using a high availability database management system (DBMS), which continuously provides a service when an error occurs in the DBMS. The DB management method in the high availability DBMS including a plurality of broker nodes and a plurality of server nodes may include an operation (a) to enable one of the plurality of broker nodes to access a main server node of the plurality of server nodes when a DB change request is received from an application server, and an operation (b) to enable the main server node to process the DB change request when the broker node succeeds in accessing the main server node, and to enable the broker node to access one of sub-server nodes so that the sub-server node processes the DB change request when the broker node fails to access the main server node due to an error in the main server node.

主权项

1. A computer implemented method of managing a database (DB) in a high availability DB management system, the method comprising:
a plurality of broker nodes and a plurality of server nodes; receiving a DB change request from an application server; in response to receiving the DB change request, enabling one of the plurality of broker nodes to access a main server node from among the plurality of server nodes; determining whether the enabled one of the plurality of broker nodes successfully accessed the main server node; if the determining determines that the enabled one of the plurality of broker nodes successfully accessed the main server node,
enabling the main server node to process the DB change request, andtransmitting a result of the DB change request to the application server; and if the determining determines that the enabled one of the plurality of broker nodes did not successfully access the main server node,
enabling the enabled one of the plurality of broker nodes to access a sub-server node from among the plurality of sub-servers nodes according to an access order defined by an access list included in the enabled one of the plurality of broker nodes,enabling the accessed one of the plurality of sub-server nodes to process the DB change request,transmitting the result of the DB change request to the application server,setting a mode of the main server node to a standby mode,analyzing a transaction log received from the enabled sub-server node, the transaction log indicating the DB change request,reflecting the DB change request in the transaction log, andwhen the main server node undergoes an error recovery process, changing the mode of the main server node from the standby mode to an active mode, the active mode indicating that the enabled one of the plurality of broker nodes is enabled to access the main server node.