Simple mysqlslap stress testing tutorial

Mysqlslap is a stress testing tool officially provided by MySQL since version 5.1.4. Simulate multiple concurrent clients to access MySQL for stress testing, and provide detailed data performance reports for "high-load attack MySQL. In addition, it can compare the performance difference of concurrent pressure of multiple storage engines in the same environment. Its syntax is as follows: shell>/usr/local/MySQL/bin/mysqlslap [Options] common parameter [Options] Details: -- concurrency indicates the number of concurrent jobs, multiple IDs can be separated by commas. For example, -- concurrency = 50,200,500 -- engines indicates the engine to be tested. Multiple engines can be separated by separators. For example, -- engines = MyISAM, InnoDB, memory -- iterations indicates the number of times a test is run in different Concurrent Environments. -- Auto-generate-SQL indicates that the SQL script generated by mysqlslap is used to test the concurrency pressure. -- Auto-generate-SQL-add-auto-increment indicates that the auto_increment column is automatically added to the generated table, starting with version 5.1.18, -- auto-generate-SQL-load-type indicates whether the test environment is a combination of read, write, update, and mixed) -- number-of-queries indicates the total number of queries to be run. -- Debug-Info indicates that information about the CPU and memory needs to be output. -- Number-int-cols indicates the number of integer attributes in the example table. -- Number-Char-cols indicates the vachar type attributes in the example table. -- Create-schema indicates the name of the custom test database. -- Query indicates a custom test SQL script. Note: during the test, a test table is generated and test data is inserted. This mysqlslap can be automatically generated. By default, a schema of mysqlslap is generated. if it already exists, delete it first. You can use-only-print to print the actual test process. After the test is completed, no trace is left in the database. Lab procedure: Exercise 1: Single-threaded testing. Test what was done.>. /Bin/mysqlslap-a-uroot-p111111 multi-thread test. Use -- concurrency to simulate concurrent connections.> ./Bin/mysqlslap-a-c 100-uroot-p111111 iteration test. Used to obtain the average value after multiple tests.>. /Bin/mysqlslap-a-I 10-uroot-p111111 Exercise 2:>. /bin/mysqlslap-auto-generate-SQL-add-autoincrement-a-uroot-p111111>. /bin/mysqlslap-a-auto-generate-SQL-load-type = read-uroot-p111111>. /bin/mysqlslap-a-auto-generate-secondary-indexes = 3-uroot-p111111>. /bin/mysqlslap-a-auto-generate-SQL-Write-number = 1000-uroot-p111111>. /bin/mysqlslap -- create-schema world-Q "select count (*) from City "-Uroot-p111111>. /bin/mysqlslap-a-e InnoDB-uroot-p111111>. /bin/mysqlslap-A -- number-of-queries = 10-uroot-p111111 Exercise 3: perform a test with 50 concurrent queries and 100 concurrent queries respectively. Execute 1000 total queries:>. /bin/mysqlslap-A -- concurrency = 50,100 -- number-of-queries 1000 -- debug-Info-uroot-p11111150 and 100 concurrency respectively get a test result (benchmark ), the more concurrency, the longer the query execution time. For accuracy, you can perform multiple iteration tests:>. /bin/mysqlslap-A -- concurrency = 50,100 -- number-of-queries 1000 -- iterations = 5 -- debug-Info-uroot-p111111 test and compare the performance of different storage engines: >. /bin/mysqlslap-A -- concurrency = 50,100 -- number-of-queries 1000 -- iterations = 5 -- engine = MyISAM, InnoDB -- debug-Info-uroot-p111111