Process restoring method and device

Abstract

The invention discloses a process restoring method and a device, wherein the method comprises the following steps: under the condition that a process of a first task is running, according to preset time parameters, taking an image file of the process as a second task corresponding to the first task and storing the image file into a designated memory, wherein the image file is used for restoring the process to a state occurring when the image file is stored; and under the condition that a first process is required to be restored, invoking the second task and running the image file so as to restore the process of the first task. By using the method and device disclosed by the invention, a process can continue to be executed from the state of the image file so as to complete the restoring of the process, therefore, the losses caused by the running collapse of users is obviously decreased, the risks of users are reduced, the problem for re-executing the process after the process fails is avoided, and an additional manual configuration required when the process is restored in different environments is also avoided; and a system is not required to carry out complicated loading, the cost of the system is reduced, the times for process re-operation and system configuration are saved, and the cost is saved.

Description

Technical field [0001] The present invention relates to the computer field, and in particular, to a process recovery method and device. Background technique [0002] In the high-performance computing environment, many user jobs need to run for a long time. In this case, the longer the user job runs, the greater the risk of user job crash and the greater the loss. If a job crash occurs during the operation of, the user can only resubmit the job and run it again. [0003] However, in practical applications, many parallel computing tasks need to modify the system configuration environment, which leads to the deterioration of the stability of the cluster computing environment. The solution to this is to manually manage the stable operation of user jobs by the administrator. Moreover, many parallel tasks are set in their own specific way, so that the process can be saved regularly, and then the task can be restarted. However, this setting is usually done manually by the administrator, ...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problems of large amount of manual operations and increased system overhead in process recovery in the related art, the present invention proposes a method and device for process recovery, which can avoid a large number of manual operations and reduce system overhead

Images