Abstract:With the development of Digital life, the distributed systems are widely used to solve a series of problems such as high concurrency. In order to ensure the accuracy and consistency of data, distributed systems need to design idempotent control. By analyzing and comparing the performance of each idempotent design method in high concurrency scenarios, an improved distributed lock design method is proposed. This method mainly solves the problem that the distributed lock may fail due to the downtime of the cluster service master node, by discussing the RedLock algorithm scheme and controversy, by discussing the RedLock algorithm scheme and controversy, propose the theory of secondary interception of high concurrent requests, and notify the invalid lock through the message queue service, realize lock failure troubleshooting and service monitoring and management. Experimental tests show that this method achieves the interception of invalid locks, has good performance in high concurrency scenarios, and provides a feasible solution for the idempotent design of distributed systems.