MySQL benchmark test, Java test

Mysql Benchmark Test

Benchmark testing is a stress test designed for the system.

First, the main work of the benchmark test:

1. Verify that some assumptions are based on the system and confirm that these assumptions are true (we do the main thing in the test, just some functional tests).

2. Reproduce some of the abnormal behavior in the system to resolve these exceptions. (such as: User login on the web often fails, 501,502 error.) ）

3. Test the operation of the current system. (System stability, bug fix, whatever).

4. Simulate a higher load than the current system.

5. Plan for future business growth.

6. Test your application's ability to adapt to variable environments.

7. Test the configuration between different hardware, software, and systems.

8. Verify that the newly purchased equipment is properly configured.

II. strategy of benchmark testing

1. Testing for the overall system

(1) It is very useful to test the entire application system, including Web servers, application code, networks, and databases, because it is not only the performance of MySQL itself that is concerned with, but the overall performance of the application.

2. Test only a single component, Mysql

(1) Need to compare the performance of different schemas or queries

(2) Testing for a specific problem in the application

(3) In order to avoid lengthy benchmarks, a short-term benchmark can be used to do a fast "cycle cycle" to detect some of the adjusted effects.

Third, test indicators

1. Throughput 2. Response time, latency (in the Web App if the response is requested, Super 1S will be looking for a problem.) ）

3. Concurrency (this is really very seldom done in development, it has to be changed.) ）

4. Scalability (basically no, consider the interface, design, these things, in the development of component applications, be aware of. ）

Iv. Benchmark test methods

1. Use a subset of real data

2. Using the error data distribution

3. In multi-user scenarios, only single-user tests are done

4. Test distributed applications on a single server

5. Does not match the real behavior

6. Execute a query repeatedly (MySQL data cache)

7. No check errors (such as a SQL that should be very slow, may be a bug)

8. Ignore system preheating (such as hard disk, at the beginning of the reading speed quickly, over a period of time, will be lowered)

9. Using the default configuration

10. Test time is too short

Only by avoiding the mistakes mentioned above can we embark on the long road of improving quality.

Read "High performance MySQL" record

MySQL benchmark test, Java test