Defined:
is to test the performance of the system by simulating various normal, peak and abnormal load conditions in a specific hardware, software, and network environment through automated test tools.
Includes: execution efficiency, resource occupancy, stability, security (stress testing is a safety test), compatibility (class files can be placed on the compatibility platform, such as the program and browser kernel can get together)
LoadRunner working principle--from Performance test interview questions
Performance Test Implementation steps
LoadRunner's scripting work
Purpose of performance testing
- Verify the improved performance effect, and need to compare with the previous test results;
- New business on-line, verify that the new system can meet the system's on-line indicators;
- Verify system stability;
- Verify that the system's architecture has bottlenecks;
Building a performance test environment
- Hardware environment: Refer to the actual production environment to build, and consider the cost of their own hardware
- Software Environment: Try to be consistent with the version and configuration used in the production environment, and keep the optimal configuration possible
- Network environment: As far as possible to reference the production environment network structure and build, as far as possible not across multiple network segments
- Preparation of basic data in a database
- The content and data volume of the underlying data
- Need to refer to the business content and usage scale of the specific system
- Scale of data volume for similar systems
- Add as much redundant data as you can to a certain percentage
- Test Data Preparation method
- Database stored Procedures
- Automated test tools such as LR, selenium, etc.
- The best performance test environment is the formal environment that will be officially launched
Performance Test Tool Selection
- LoadRunner
- Webload
- Jmeter
- Self-writing multi-threaded pressure test program
Selected criteria: Performance testing requirements that fulfill specific requirements, prioritizing LR with powerful charting and control functions
Definition of performance test scope
- Identify the scope of performance tests with the entire development team
- Frequently used functions in the system, interfaces for calls, etc.
- The system involves a large number of database read and write functions
- The ability to read and write large portions of the system cache
Performance Test Requirements Acquisition
- Http://www.cnblogs.com/scios/p/5432599.html
Completion targets for performance tests
- The new on-line test system does not have a clear digital standard contrast, the system has been tested to the system limit (some of the system's resources have been exhausted, CPU, handle, memory, the database has a large number of slow query, the system has been slowed down some processing), and the system proved to be able to scale the level, You can go online.
- Compared with previous test results, the test results are better than previous test results (TPs more, faster processing of single requests), as long as the test conditions are proved to be similar.
- There is no comparable test results, but the product has been online for some time (at least three months), some operational data, you need to analyze the operational data to be used as a benchmark, as long as the system reached three months in the system concurrency Peak 4 times times can be considered acceptable. (If the interface is a test object, you need to mix the primary interface for performance testing)----not recommended
- If the developer provides experience as a benchmark, the object being tested can only prove to meet the developer's experience value.
- If you choose one of these strategies, you must specify the number of processes per second and the exact time of each request, and issue a final performance test report.
Script Debugging for performance testing
Script recording, writing, modifying, randomization policy, debugging and running, viewing log and database content validation script correctness
Data collection for performance testing
Performance test Data analysis
Experience Introduction
Training ideas for performance testing