This blog covers different aspects around software development and test. There are also general information useful for who needs to test performances and needs to analyze bottlenecks on application servers and databases. Disclaimer: test before you implement any advice, no warranty is provided. Before use the software mentioned in this blog read and understand all licenses agreements. Use of the information contained in this blog is under your own responsibility.
venerdì 23 agosto 2013
How can increase the data population to simulate a medium-high load condition?
If your web application is database centric, when you need to test performances you will start thinking how.
How can increase the number of items on the tables to simulate a medium-high load condition?
First don't do it manually!
Let's see some options:
1. Take a tool that has the capability to record and playback, start recording the scenario or all the scenarios necessary to create the data you need. Then customize the scenarios providing parametric values for each input field of the interface, then playback the scenarios how may times you need, so each time the scenario automatically run creates the data.
2. Run all user scenarios, enable debug logs on your application or on the database and sniff all sql statements (insert, update) executed during the scenarios. Then write some piece of code, in any language you prefer, using those sql statements and customizing the input values as it is necessary for your test. Finally run the program and populate how much you need. I've used IBM Rational Performance Tester and can do all this.
The first approach has the advantage of creating the data with a blackbox approach, so it is the web application itself that creates the data not you. So you cannot do any error. But the disadvantage is it could take very long time to insert many items.
In the second approach you cannot consider the database as blackbox and you need to have some experience with database programming, knowledge of the database schema and relationships. You have to make sure that all tables involved on the scenarios have been considered. You can make errors. In some case it needs to check on the design documents to validate each types and values and relationships. The big advantage of this approach is that you can insert very big data and simulate the load of the most exigent customer scenario.
Iscriviti a:
Commenti (Atom)