基于PBS的勘探数据处理作业管理

PBS-based Job Management for Exploration Data Processing

高性能集群不具备作业自动调度和负载均衡的功能。采用开源的作业管理系统定制开发集群管理系统,解决集群"难用难管"的问题。作为一种开源的队列管理和作业调度系统,PBS目前已经广泛应用于集群管理当中。通过Shell脚本应用开发,将不同类型的应用作业转换为相应的PBS作业脚本纳入系统管理。利用PBS系统进行必要的定制开发工作,在较少改变科研人员工作习惯的前提下,实现Paradigm公司EPOS处理系统集群队列管理和作业分发管理。

High performance cluster does not have the functions of job automatic scheduling and load balancing. We use open source job management system to develop cluster management system in order to solve the problem that the cluster is difficult to use or manage. As an open source queue management and job scheduling system, PBS has been widely used in cluster manage- ment. Through developing a number of Shell scripts, the different types of jobs are transformed into the corresponding PBS job scripts. Based on the actual situation of work mode in the enterprise cluster and the premise of less changing scientific research personnel work habit, by using the PBS system to make the necessary custom development work, the Paradigm company EPOS system cluster queue management and iob distribution management are imvlemented.