1) Is it possible to test multiple servers at the same time (single client machine)

   You can do with multiple servers. Please find the details below:

If you are asking about configuring application with more than one app/web servers and measuring the performance then yes we can measure the performance of the application. Executing the loadrunner does not bother about the application configurations and it just replicate real user. Ex: Considering the real user accessing to the web application, he does not bother about the configuration but he expects proper response to the request he made at application. Consider the application configured with 1 web servers , 3 app servers and load balance enabled. In this case loadrunner speaks to web servers through the application URL used to create LR scripts. We need to know the architecture of the deployment but only need to record script and collect required stats of all serves configured to the application for analysis to identify bottlenecks.

{mosgoogle left}  

2) How to use controller and analyzer (real scenario)

          Controller and Load Generator

   The LoadRunner Controller is the console for managing the entire load-testing scenario. It runs on NT/Win2K machines only and performs the following tasks:

•       Controlling and monitoring Driver machines, scripts, and Virtual Users.

•       Monitoring Network and Server statistics during tests

•       Displaying messages and status of users during tests

•       Collecting performance results.

Close physical proximity between the Controller and Host machines is not necessary, however, a LAN connection must exist between them.

 Creating a Scenario

Create a scenario in order to test a system.  A LoadRunner scenario is a file that contains information about the entire testing session.  More specifically, it includes:

•       The groups of virtual users (Vusers)

•       The test scripts that each Vuser will run

•       The load generator machines upon which to run the scripts.

Create a list of available host machines and scripts. Then create Vusers or Vuser Groups (collections of Vusers performing similar tasks) with the corresponding scripts and host machines that you want them to use.

The Controller displays the host, script, and Vuser lists. To view one of the lists, click on the appropriate tab.

{mosgoogle left}  

        A script defines the actions that a Vuser performs during scenario execution. The script list contains all the Vuser scripts that Vusers can run. The script list may    contain script for all types of Vusers.

Analyzer:

       Analyzer is LR utility/tool used to view the LR results captured in readable format. Ex: Analyzer converts the stats collected to graphs, averages the response times collected..etc. User can filter the results data for particular duration of time which helps to identify the root cause.

3) Important parameters for analysis in load tests.

•       Transaction response times.

•       Vuser status

•       Throughput

•       Servers(web,app,Db..etc) performance stats

•       Network latency

4) Usual statistics collected out of loadrunner

  Find the attached document which gives more info about the possible monitors can be collected from loadrunner. Some of the monitors to add we need access permission's or ports open behind the firewalls to collect stats. Ex: unix server/DB stats to collect.

5) Can issues be categorized at different tiers of application (application, database...)

   ---Yes. For better and in depth analysis issues should be categorized by isolating the stats specific to each layer. Ex: If we have A,B,C,D are different tiers/layer considering A as web, B&C as app, D as DB. At particular duration of time if we notice the transaction time are increasing from LR graph then collecting the individual server stats will helps to analyze the server affecting the performance. We need to collect network stats between each tier to analyze the exact network bottleneck.

6) How to capture dynamic values in a web application

------If you are asking about capturing dynamic values in web application using correlation then the following steps gives brief about easy way of manual correlation. This method is time consuming initially but effective approach.

    1.  Record an identical script using a different user.

    2.  Compare the two scripts for differences to determine dynamic values.

    3.  Use the functions, web_create_html_param and web_reg_save_param to create parameters to hold the dynamic values.

    4.  Replace hard-coded values in the script with parameters.

    5.  Verify correct execution of the script by replaying the script

7) How should the scripts be structured. (samples will help. can remove any specific information).

Aim to easy maintenance of scripts

If you mean maintenance of script is modifying the scripts according to the minor modification of the application/product then it is different with LR scripts. LR scripts are generated based on the back-end requests/responses, so most of the application changes with respect to logic will affect the scripts. In this case re-recording the scripts is best option to capture the proper request/responses which helps to capture accurate measurements. Please let me know with more details for additional info.

8) Is it possible to send results to QC

We can make connection to QC and save the results. But i have not worked on it so cannot provide much info.

9) Script execution from Quality center

10) How are correlations used