One kind of geological data based on the mass distributed file system parallel processing method

The present invention discloses a mass geological data based on the distributed file system parallel processing. The present method is: 1) distributed file system as a storage system geoscience data, deployed on a computing cluster; distributed file system with a unified namespace; save computing jobs job scheduling system 2) computing clusters to a received waiting queue; 3) scheduling system selected from a waiting queue into the run queue computing jobs; 4) scheduling the computing system computing node to find the data required to run the job file is located in the metadata in a distributed file system in accordance with the job information calculated ; 5) holding a job scheduling system pick up the computing node data required for the operation of the computing job; computing node remote acquisition of the computing job but need not hold the computing node data file, and the calculation is performed in the computing node job and return the results. This method of computing the maximum extent do localized.