Distributed system exception unified processing method

Abstract

The invention relates to a distributed system exception unified processing method which comprises the following steps: A, establishing a calling process interceptor class based on a serial interface for initiating remote calling, capturing exceptions and printing related logs; b, establishing a result response class, and responding to a calling result among the services; c, establishing a link tracking tool class based on Thread Local, RpcContext and MDC for setting a request number; d, the service party and the consumer respectively configuring the calling process interceptor class in an XML mode; e, establishing a consumer tangent plane interceptor class for intercepting before the consumer calls the method; and F, establishing a service party tangent plane interceptor class for intercepting before the service party calls the method. According to the method, the development cost is effectively reduced, abnormity can be rapidly positioned and analyzed, and unnecessary risks caused by internal code confusion are prevented.

Description

technical field [0001] The invention relates to a method for unified processing of abnormality in a distributed system. Background technique [0002] In the processing of program exceptions in distributed systems, the DUBBO RPC framework provides basic exception handling components. In the case of exceptions, custom exceptions can be thrown to consumers, and consumers need to understand each exception type and target Different exception types make different handling strategies. If the consumer does not show the exception to be caught, it will cause a fatal error to the business system. The official documentation of the DUBBO framework can be referred to: http: / / dubbo.apache.org / zh-cn / docs / user / quick-start.html [0003] It can be seen that the shortcomings of the existing exception handling methods are: [0004] 1. When the server defines the API interface (application programming interface), it needs to clearly define various exceptions that may be thrown on the interface,...

AI Technical Summary

Problems solved by technology

This undoubtedly increases the workload of the service provider

[0005] 2. The consumer needs to capture each exception separately and make different handling strategies for the exception. If the exception is forgotten or missed, it will cause a fatal error to the business

Therefore, the displayed exception capture also increases the workload of the consumer

[0006] 3. The conventional method of throwing exceptions by the system is not friendly to consumers and has performance problems

[0007] 4. In the case of distributed calls, multiple service parties are likely to be called in order to complete a business entry. When an exception occurs on the business call link, the existing method is difficult to quickly locate the call link context and Quickly locate the cause of the problem

Images