CN106775974B

CN106775974B - Method for realizing distributed priority queuing lock

Info

Publication number: CN106775974B
Application number: CN201611118040.1A
Authority: CN
Inventors: 姚文俊; 杨松; 季统凯
Original assignee: G Cloud Technology Co Ltd
Current assignee: G Cloud Technology Co Ltd
Priority date: 2016-12-07
Filing date: 2016-12-07
Publication date: 2019-12-10
Anticipated expiration: 2036-12-07
Also published as: CN106775974A

Abstract

the invention relates to the technical field of distributed concurrency and JAVA development, in particular to a method for realizing a distributed priority queuing lock. The invention firstly sets a priority for the request which needs to execute the operation; and then starting to acquire the lock, judging whether the mutual exclusion field in the database is empty, if not, failing to acquire the lock, otherwise, judging whether a waiting object exists in the queue or not, and whether the first bit in the queue is the self or not, if so, setting the mutual exclusion field, and otherwise, failing to acquire the lock. If the setting of the mutual exclusion field is successful, the lock is successfully acquired, and if the request processing is completed, the mutual exclusion field is released; and when the lock acquisition fails, if the request is not in the queue, the priority is raised, if the priority is raised to a preset value, the request is added into the queue, and the steps are repeated until the lock acquisition is successful or the time is exceeded. The invention solves the problems that the request can not obtain the lock, the lock obtaining efficiency is not high, and the priority is not flexible, and can be used in the distributed environment with large concurrency and needing to process the specific request preferentially.

Description

method for realizing distributed priority queuing lock

Technical Field

The invention relates to the technical field of distributed concurrency and JAVA development, in particular to a method for realizing a distributed priority queuing lock.

background

The method of using lock is generally used when processing data exception caused by concurrent request of system, and in distributed system, it is generally realized by memcached or redis distributed lock. However, when the concurrency is larger and the request situation is more and more complicated, the existing scheme has the following problems:

firstly, the solution of the existing distributed lock is preemptive, and who acquires the lock first and who handles the lock first. In addition, when the concurrency is large, a request can be processed for a long time due to the fact that the lock cannot be seized, and even the request cannot be processed due to the fact that the lock cannot be seized all the time;

secondly, the problem is solved by a queuing lock commonly used on the existing single server, but the queuing lock is not realized on a distributed system, so that the problem cannot be solved by the distributed system, and meanwhile, if the queuing lock is directly used, the lock acquisition efficiency is low because the lock must be acquired in sequence;

thirdly, the existing lock scheme cannot set the priority, so that the flexibility is low, the requests needing to be processed preferentially cannot be processed preferentially, and the requests can only be preempted or queued together;

in order to ensure the lock acquisition efficiency, prevent the problem of overlong lock acquisition time and improve flexibility, a method for continuously increasing the priority and customizing the priority configuration after the lock is unsuccessfully seized in the lock acquisition of the distributed system is required.

disclosure of Invention

The technical problem to be solved by the invention is to provide a method for realizing a distributed priority queuing lock, which can ensure the lock obtaining efficiency of the concurrent request in a high-concurrency environment of a distributed system, solve the problem that the lock can not be obtained by the request and provide a more flexible priority to meet various special requests.

the technical scheme for solving the technical problems is as follows:

the method comprises the following steps:

Step 1: setting a priority and a priority strategy for each request, and starting to acquire a lock;

step 2: judging whether the exclusive field of the database is empty, if so, executing the step 3, and if not, executing the step 6;

and step 3: judging whether the queue has a waiting request, if the queue has no waiting request, executing the step 9, and if the queue has a waiting request, executing the step 4;

And 4, step 4: judging whether the request is in the queue, if so, executing step 5; otherwise, executing step 6;

And 5: judging whether the request is at the first bit, if so, executing step 9; otherwise, after waiting for a period of time, re-executing the step 2 and continuing to acquire the lock;

step 6: raising the priority policy of the request according to the priority policy;

And 7: if the priority reaches the threshold, executing the step 8 if the priority reaches the preset priority threshold, otherwise, waiting for a period of time, executing the step 2 again, and continuing to acquire the lock;

And 8: putting the request into a queue, re-executing the step 2, and continuously acquiring the lock;

and step 9: setting a mutually exclusive field of the database, executing the step 10 if the setting is successful, and executing the step 6 if the setting is failed;

Step 10: the request is processed and the database mutual exclusion field is cleared and the process ends.

The priority strategy is to set the step length of priority increase after the failure of the mutual exclusion field or calculate a new priority according to a formula;

the mutually exclusive field of the database means that the field can only be set and updated by one request, and once one request is successfully operated, other requests cannot be successfully operated.

the queuing queue is a queue of requests in which the requests get the lock in order, and the system will prioritize the requests in the queue.

The threshold is a preset priority value, and when the priority of the request reaches the value, the request can enter the queuing queue.

The lock acquisition means that a token is acquired, and the lock acquisition can be executed only when the token is taken, so that data errors caused by concurrent processing of requests are avoided.

The invention develops a queuing lock method capable of flexibly setting priority by combining the advantages of the traditional distributed lock in a distributed system, so that the distributed system can simultaneously request for maintaining the lock acquisition efficiency, avoid the problem of long-time waiting of individual requests, simultaneously can promote the priority of the individual requests and meet special requirements. By solving the problems, the invention can improve the request experience of the distributed system, improve the satisfaction degree of users, optimize the system, avoid the request blockage and reduce the operation efficiency of the system.

Drawings

The invention is further described below with reference to the accompanying drawings:

FIG. 1 is a flow chart of the method of the present invention.

Detailed Description

As shown in FIG. 1, the process of the present invention is as follows:

main flow:

releasing the lock:

queuing to obtain the lock:

updating the priority and judging whether to join the queue:

Setting the priority:

。

Claims

1. A method for realizing distributed priority queuing lock is characterized in that: the method comprises the following steps:

2. the method of claim 1 for implementing a distributed priority queuing lock, wherein: the priority strategy is to set the step length of priority increase after the failure of the mutual exclusion field or calculate a new priority according to a formula;

3. the method of claim 1 for implementing a distributed priority queuing lock, wherein:

4. The method of claim 2 for implementing a distributed priority queuing lock, wherein:

5. A method for implementing a distributed priority queuing lock according to any of claims 1 to 4 wherein:

the threshold is a preset priority value, and when the priority of the request reaches the value, the request enters the queuing queue.

6. A method for implementing a distributed priority queuing lock according to any of claims 1 to 4 wherein: the lock acquisition means that a token is acquired, and the lock acquisition can be executed only when the token is taken, so that data errors caused by concurrent processing of requests are avoided.

7. the method of claim 5 for implementing a distributed priority queuing lock, wherein: the lock acquisition means that a token is acquired, and the lock acquisition can be executed only when the token is taken, so that data errors caused by concurrent processing of requests are avoided.