| 摘要 |
A method of detecting a fault or a potential fault of a DBMS operating in a clustered environment is provided. The DBMS includes a log of diagnostic information that is related to the status of the DBMS. At predetermined intervals, monitoring of the log of diagnostic information is conducted for at least one of a plurality of predefined conditions occurring anywhere in the clustered environment. The predefined conditions are related to a fault or a potential fault of the DBMS. Before a fault of the DBMS occurs, the presence of at least one of the predefined conditions in the log is detected. A response to the detected predefined condition (e.g., a fail-over) is provided.
|
