| 摘要 |
A method and apparatus that provides informed prefetching and enhanced memory utilization in a server cluster to improve failover and startup time of a resource group executed by a server cluster. When starting up or failing over a resource group on a given server, the method identifies information for prefetching, informs a file system executed by the server cluster of the identified information, accesses the identified information using the file system, and stores the identified information in main memory for use by the server in the server cluster in advance of when the server actually will need the identified information to failover or initiate resources. To enhance memory utilization the method also identifies pages in cache memory that are not needed for current services provided by a server. Generally these pages contain prefetched information that was already used at start up or failover of a resource group. The identified pages are placed on top of a free page list such that those pages will be reused prior to any other pages in the list being reused. |
