Critical resource management method and device in cloud storage system

Abstract

The invention discloses a critical resource management method and a critical resource management device in a cloud storage system. The critical resource management method comprises the following steps: when a process x needs to apply to occupy a critical resource y, establishing a session connection with a managed node; after successful establishment, initiating a request of occupying the critical resource y to the managed node; utilizing a predetermined value to replace the value on the flag bit corresponding to the stored critical resource y by the managed node; if the replaced value is identical with the predetermined value, then determining that the critical resource y is occupied and the application fails, and notifying the process x; or, determining that the critical resource y is not occupied and the application is successful, and notifying the process x. With adoption of the scheme, the implementation complexity can be reduced, and the stability of the system performance can be improved.

Description

technical field [0001] The invention relates to storage technology, in particular to a critical resource management method and device in a cloud storage system. Background technique [0002] In cloud storage systems, applications for critical resources are generally done through distributed locks, which evolve from stand-alone locks. The cloud storage system refers to a system that integrates a large number of storage devices in the network through application software to work together through cluster applications, network technology, and distributed file systems to jointly provide data storage and business access functions. [0003] However, the above method has certain problems in practical application, such as: [0004] For each process (locking process) that applies to occupy a critical resource, a corresponding notification interface needs to be set. When a critical resource is released (unlocked), all processes that apply to occupy the critical resource need to be not...

AI Technical Summary

Problems solved by technology

[0004] For each process (locking process) that applies to occupy a critical resource, a corresponding notification interface needs to be set. When a critical resource is released (unlocked), all processes that apply to occupy the critical resource need to be notified through the notification interface , so that these processes select a process to occupy the critical resource through competition, etc.; it is more complicated in terms of implementation;

[0005] In addition, for each critical resource, a corresponding waiting queue will be set up, and each process that applies for occupying the critical resource will be added to the waiting queue for waiting. In this way, when the critical resource is released, the waiting queue All processes will be awakened to seize this critical resource, and only one of them will succeed in preempting, and the rest will continue to wait. Usually, the phenomenon of a large number of processes being awakened at the same time to seize resources is called the shocking group phenomenon. The appearance of the shocking group phenomenon will cause system The performance fluctuates sharply, which leads to extremely unstable system performance

Images