| 摘要 |
A method, apparatus, and article of manufacture for resolving long-busy conditions for synchronized data sets. A long-busy condition is detected in a first one of the synchronized data sets. Further access to the first data set is prevented and all subsequent accesses are directed to another, available, second one of the data sets during the long-busy condition. All updates that are made to the second data set during the long-busy condition are identified and stored in a data structure. An untimed read operation is used to determine when the long-busy condition has cleared. Once the long-busy condition is cleared, the first data set is placed into a recovery mode, wherein the identified updates are applied to the first data set. During the recovery mode, the first data set is read-inhibited to prevent read accesses thereto, but the first data set is write-enabled to allow write access thereto. Upon completion of the recovery mode, the first and second data sets are re-synchronized and normal processing can resume.
|
