Scenarios
A scenario contains all the settings necessary to run a load test:
Scenarios
NeoLoad can handle several scenarios by project. By this way, it is easy to switch between several kinds of test:
a short one with a ramp-up load, a very long one with load peaks and a one with a very realistic scenario....
The load simulated in a scenario is defined by duration and by a number of Virtual Users.
Each virtual user simulated plays its web pages definition then ends. Then, NeoLoad launches a new Virtual User of
the same kind to keep the load at the desired volume.
Set the duration of the test: unlimited, a specified amount of time or by iterations.
One iteration represents one Virtual User complete playback. When choosing the duration by iteration, each
population execution duration must be defined by setting the number of iteration in the Load Policy panel.
In all cases the test can be stopped at any time by clicking the "stop test" buttons.
Specify the number and the variation of Virtual Users to generate.
Choose one or more populations to test and each population's load policy:
- Constant - Generates a constant number of Virtual Users.
- Ramp up - Generates a growing number of Virtual Users.
Useful to see how the server behaves when the load increases.
- Peaks - Generates a constant number of Virtual Users with periodical phases of high load.
Useful to see if the server retrieves a normal behaviour after a peak of high load.
In iteration mode:
The population will be played for a iterations number defined in this panel. Each simulated user is launched for
the specified number of iterations and doesn't wait between 2 iterations. When the end of the test is near, some
users have finished all their iterations, others are still playing. This can explain that the load is falling down
until all users are terminated.
For the Ramp up and the peak, the launching of new users after a volume step is made when all previous users have
terminated their iterations. So before each step, depending on the definition of virtual users, the load
can be falling down near 0.
- Constant - Define the number of iterations that will be played by each virtual user.
- Ramp up - The number of iterations defined is used for the initial simulated users.
For example, consider a ramp-up from 10 users, with increment of 2 users every 5 iterations. Setting
an iteration number of 20 will give a test like:
- 10 users for 5 iterations
- 12 users for 5 iterations
- 14 users for 5 iterations
- 16 users for 5 iterations and then end.
- Peaks - The number of iterations defined is used for the initial simulated users.
For example, consider a peak from 10 users as minimum load for 5 iterations and
20 users as maximum load for 3 iterations. Setting an iteration number of 20
will give a test like:
- 10 users for 5 iterations
- 20 users for 3 iterations
- 10 users for 5 iterations
- 20 users for 3 iterations
- 10 users for 4 iterations and then end.
The Load Generator is the program that simulates the Virtual Users.
Each NeoLoad controller contains a Load Generator. Load Generators can be installed as separate programs on different computers.
Load Generator Hosts lists the registered Load Generators. All the Load Generators running in the local network are automatically detected
on NeoLoad Controller start up. Press "Discover" to auto discover Load Generators started since (May need a few seconds to be detected).
Load Generators can be manually declared: enter hostname or IP address.
Running Load Generators get a green sign while undetected ones get a red sign.
Choose the Load Generators used to play the test. The controller will distribute the Virtual Users to simulate among
the selected active Load Generators.
Advanced Host Configuration
Ip Spoofing tab
All the detected Network Interfaces of the LoadGenerator are listed here. For hosts acting as router on several networks, you can choose the
network card used by the LoadGenerator.
On the selected interface, all the IP addresses of the host are listed.
By default, when only one IP address is selected, all virtual users generated by the Load Generator Host use this IP address.
Defining multiple IP addresses allows the testing of applications behind a load balancer based on the IP address.
In that case, a randomly chosen IP address is assigned to each Virtual user.
The IP addresses must be available in the network i.e. not used by other computers of the network. Ensures theses addresses can be used with your administrator.
To define IP addresses in the list, you have to define additional IP addresses in the Load Generator Operating System settings.
Operating System configuration
Theses settings will modify the network configuration of the Load generator. Please consult your administrator before modifying the host configuration.
The multiple IP addresses settings are also called virtual IP settings in some documentation.
- Windows
- Step 1: In the Start Menu, click on the Control Panel.
- Step 2: Open the Properties of the Local Area Connection: right-click on the Local Area Connection and click on Properties.
- Step 3: Select Internet Protocol (TCP/IP) in the list then click on the Properties button.
- Step 4: The multiple addresses settings are not available in DHCP. If the DHCP configuration is selected i.e. "Obtain an IP address automatically" is selected, you will have to change it to a fixed address.
- Step 5: The IP addresses panel shows the list of all defined addresses. Click on the Add button.
- Step 6: Enter the new IP address and the subnet mask in the dialog box. Then click on the Add button.
- Step 7: The new IP address should appear in the list.
- Repeat from step 5 to step 7 to define each IP address.
- Linux
The multiple addresses settings are not available in DHCP. Set up your configuration to use fixed IP addresses.
Use the command ifconfig to add a new IP address.
For example, the command lines to add 2 new virtual IP addresses (ex: 192.168.1.10 and 192.168.1.11) to the network interface eth0 are:
- ifconfig eth0:0 192.168.1.10
- ifconfig eth0:1 192.168.1.11
Control that all the ip addresses are well added with the command: ifconfig
Report to the IP Aliasing Mini-HOWTO to get more information.
Load Balancing tab
The virtual users simulated in the test are distributed among the available Load Generators.
By default, all the Load Generators create the same number of virtual users.
By increasing the load factor, you increase the number of virtual users that will be created on this Load Generator.
For example, if you have a Load Generator "A" with a factor of "3" and a Load Generator "B" with a factor of "2",
"A" creates 3 virtual users while "B" creates 2 virtual users.
The most powerful computers should be parameterized to generate more Virtual Users.
Advanced
Allows defining:
- The start policy - Defines how to start the Virtual Users: simultaneously or within a delay.
This policy is used each time new Virtual Users are created: on each load increment for a Ramp-up load policy
and on each peak for a Peak load policy.
- The Thinktime policy - Defines if NeoLoad should use the thinktime of each page or a common thinktime for all pages.
Use a Random Integer variable to get a random value.
Declare and use performance counters. These counters allow monitoring any Windows machine during the test, tracking any issues
on the web server, application server or database server.
Pick and rename Windows performance counters.
Note: to access a remote computer, you must log on as Administrator or as a member of the Administrators group. If the computer on which the controller is running is not logged with the Administrator account, launch the controller with the appropriate account by using the "Execute as..." option of the right click menu on NeoLoadGUI.exe.
On a typical server, important basic counters are:
- Memory: Pages/sec - Detect swapping: Pages/sec is the number of pages read from or written to disk to resolve page faults.
A page fault is when a process needs to retrieve code from the physical disk because it is not available in memory.
- Processor: \% Processor Time - Percentage of CPU activity.
- Physical disk: \% Disk Time - % Disk Time is the percentage of elapsed time that the selected disk drive is busy servicing read or write requests
- Network: Bytes Total/Sec - Bytes Total/sec is the rate at which bytes are sent and received on the network interface
When declared, the performance counters are available for all scenarios within the same project.
Select which counters to use for each scenario.
Unix counters will be supported soon.
© Copyright 2005, Neotys.