摘要： 提出了一种两级元数据服务器机群文件系统的负载平衡策略，高级元服务器依据与二级元服务器性能最密切相关的负载信息将任务快速分配，在存储文件时，提出了一种能正确反映各存储节点I/O流量和存储量的方法：计算文件热量值，并据此对待存文件进行了合理的分配存储。实验结果表明，该策略提高了系统的性能，缩短了任务执行时间，取得了较好的效果。

关键词: 机群文件系统, 元数据服务器, 负载平衡, 热量值

Abstract: This paper explores a load-balance algorithm for a cluster file system which contains two levels of metadata-server. The primary-level-server quickly distributes tasks to second-level-server depending on the closest load-balance information. A method which accurately reflects I/O traffic and storage of storage-node: computes the heat-value of file, on which a more logical storage allocation is realized. Experimental results show that the algorithm shortens the executing time of tasks and improves the system performance compared with other load algorithms.

Key words: cluster file system, metadata-server, load-balance, value of heat

中图分类号: 

• TP393