Method and device for treating abnormality of task flows

Abstract

The invention relates to a method and a device for treating abnormality of task flows. The method comprises the following steps of: (1) converting a user request into a task flow consisting of a plurality of task nodes, and performing the task flow; (2) judging whether the task nodes are performed abnormally or not; (3) temporarily stopping performing the task flow if a certain task node is performed abnormally, and placing the information of the abnormal task node into multi-level waiting queues which respectively indicate a plurality of different waiting time; and (4) performing the abnormal task node according to the information of the node. By the method and the device, the request can be performed completely, the final integrity and the consistency of the data and the state of the request are guaranteed, manual recovery efficiency is improved, and a uniform and stable interface is formed.

Description

technical field [0001] The invention relates to a method and device for processing abnormal task flow. Background technique [0002] Task flow is a background processing process of website business. In the business system of the website, there are some complete large tasks composed of multiple subtasks. These subtasks are only used to complete a certain kind of work, and these subtasks are highly reusable, so these subtasks cannot be completely integrated. Subtasks are integrated into a larger task. In order to complete such a large task, it is necessary to define some task flows, connect all subtasks, and execute them to completion. Taking the completion of message communication requests on e-commerce websites as an example, this major task includes: the sender sends some in-site messages to the recipient, the sender sees the message in his sending record, and the recipient sees the message in his See the message in the inbox of the sender, that is, from the time the sen...

AI Technical Summary

Problems solved by technology

For an e-commerce website, the task flow often involves the transaction data of the user, and such processing is obviously unreasonable; moreover, in the case of deleting half of the inserted data, although it can ensure that the data is consistent as a whole, However, it still does not respond correctly to the completion of the request, which may easily cause the user's important data to be discarded.

[0006] 2. When the system fails to process a certain task flow in the background at a certain moment, a single manual recovery method is adopted, which cannot timely and targetedly respond to various abnormalities (for example, system dirty data, system logic error, requested error, etc.) Some conditions have not yet been met, etc.) to be treated differently

[0007] 3. Manual recovery is cumbersome, error-prone, and inefficient: For all error scenarios, manual recovery is used. Although there is hope for data recovery, it is inefficient to do so, and in most scenarios, human intervention is not necessarily required. Participation is required to recover the data; in addition, when the amount of manually recovered data increases, the error rate increases and other types of problems arise

[0008] 4. The interface is inconsistent and unstable due to simple abandonment of requests or deletion of data

Images