Method and device for testing performance of database

Abstract

The invention provides a method for testing performance of a database. The method comprises the steps that a pressure model generating module generates pressure test data according to pre-generated SQL statements; more than one thread is adopted in a pressure execution module to send the generated pressure test data to a database cluster; a monitor module collects the performance information of the database; an analysis module judges whether the collected performance information of the database is located in a preset reasonable range or not, and a first test analysis report is generated according to a judgment result. The invention further provides a device for testing the performance of the database. According to the technical scheme, the performance of the database can be tested conveniently and flexibly.

Description

【Technical field】 [0001] The invention relates to testing technology in the Internet field, in particular to a database performance testing method and device. 【Background technique】 [0002] All Internet operators are faced with the problem of database type selection and database version selection. To solve these problems, it is necessary to rely on the performance test of the database and obtain a large amount of real test data. In addition, database and application development also require database performance testing to eliminate database bottlenecks. [0003] At present, the industry generally uses testing tools provided by third parties to directly test database performance, for example, testing tools such as Sysbench, Supersmach, Benchmark SQL, and Jmeter. In this database performance testing method, there are the following problems: [0004] 1. The Structured Query Language (SQL, Structured Query Language) statement used for testing is not obtained from actual data,...

AI Technical Summary

Problems solved by technology

[0004] 1. The Structured Query Language (SQL, Structured Query Language) statement used for testing is not obtained from actual data, and it is impossible to simulate the actual scene to obtain test data similar to the actual data. The simulation is poor, resulting in inaccurate test results

[0005] 2. The test tool provided by the third party has a complex framework and many components, so the operation is complicated when installing the test tool, requiring a lot of manual deployment, and the portability is poor

[0006] 3. The test rules of the test tool provided by the third party are fixed, and the operator cannot carry out secondary development of the test tool for different services, so the test tool provided by the third party cannot be flexibly applied to the database test of various services

[0007] For example, when using the test tool Sysbench to test the performance of the database, the installation of Sysbench needs to rely on make and the lib library of MySQL, and it cannot implement multi-process testing on a single machine

Although you can specify the database for testing, you can only send SQL pressure files with fixed rules; and even if you can configure the engine type, concurrent number, submission mode, and read-write type, you cannot specify specific business-related SQL statements. It can only realize the online transaction processing system (OLTP, On-Line Transaction Processing) test, and cannot simulate the actual data in the actual scene

Moreover, when testing with third-party testing tools, each test is an independent behavior, and it is impossible to correlate the test results and guide future testing operations.

Images