Shulou(Shulou.com)06/02 Report--

Sysbench 1.10 why the original command is invalid, I believe that many inexperienced people do not know what to do, so this article summarizes the causes of the problem and solutions, through this article I hope you can solve this problem.

Recently, a large number of MYSQL of the company are going online, so it is unreasonable not to do stress tests, so take out the sysbench that has been used all the time to stress test MYSQL, and the problem begins, the earliest one is 0.5 version.

Various commands are familiar, and the latest version of sysbench is 1.10.

Students familiar with version 0.5 of sysbench may remember the command set, prepare, warmup, run, cleanup, and can adjust the number of tables tested, the amount of data inserted in the table, and so on. It is relatively easy to use, and finally can give you an index of the overall QPS TPS of the MYSQL you are testing, etc., and during the stress test, you can also observe the overall pressure of the system, at least one production server is online, you need to know the bearing capacity of your server and so on.

Students who may be accustomed to other database servers generally go online instead of stress testing. The following figure shows that during the stress test, MYSQL reported an error message, while at this time, innodb_buffer still has a lot of unused memory. When you get the data such as CPU, memory and Imax O at the time when the error is reported, you can get the alarm point and default value of the system in the system monitoring. When the monitoring system reaches such a value, you should do something, but most of them seldom do stress tests.

The problem is that this 1.10 version of sysbench is compared with 0.5, both the parameters that can be manipulated and the result indicators that can be obtained are very different from before, and the relevant information from baidu google also stays in the old version of the method of operation, the new version is basically I did not find, detailed instructions, and the information on github is also very "stingy". Is unable to find of course the pressure test of the sound dripping.

I have no choice but to study it by myself, so I have this article.

From the sysbench source code, I can't find the information about where to configure table and table size, as well as some previous configuration information.

When using the new sysbench test, I found some problems, that is, even if I give the values of thread, event, etc., it basically doesn't change the middle result of my test. It's just that there are more threads.

Finally, through the lua script that comes with each command, I figured out how to control the amount of data that generated the table as before.

After installing sysbench, (compile and install), you should be able to get a series of scripts for lua in the system directory. If you open a lua script at random, you can clearly see that there is a need for oltp_common, and among the many scripts, I have not found any information about table size and how many tables.

Follow the vine and find oltp_common.lua.

The related item settings are originally here, changing the SIZE and the number of tables in the lua, as well as some other settings, such as how many transaction are merged at a time, and so on.

After reading the above, have you mastered the method of why the original command is invalid in sysbench 1.10? If you want to learn more skills or want to know more about it, you are welcome to follow the industry information channel, thank you for reading!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.