| 摘要 |
A resource coordination method, device, and system for a database cluster. The method comprises: an active coordination node acquires state information corresponding to each computing node in multiple computing nodes (201), the state information being used to indicate the running load state of the computing node; determine, according to the state information corresponding to each computing node in the multiple computing nodes, whether the active coordination node has idle resources with a preset threshold X (203); and if yes, instruct each computing node to upload clean page data subsequently generated to the active coordination node (205). By means of the method, more data generated by a computing node can be stored in a coordination node, and an idle resource on the coordination node and a feature of high speed communication between communication interfaces of the coordination node and the computing node are fully used, so that the computing node is started with a higher speed and higher efficiency. |
