Alfresco Benchmark Load Testing Login with Different Users Randomly On Alfresco Share

327 views Asked by At

I have downloaded Test Alfresco-benchmark-server and Alfresco-benchmark-test-share application and I have to load test the application while creating the test cases from the alfresco-server home page(http://localhost:8080/alfresco-benchmark-server/). Scenarios is I have to give more than one number of users in Share load parameter and when I run the test case two different user must login on the alfresco site, and their dashboard will open on two different instances of the provided browser.

But this is not happening when I run test cases even if I provide 3 number of users, and 3 session count, same single user gets login on three different instances of the provided browser. As WebDroneStratEventProcess class is getting only one user in the data object, it must get different random users from the MongoDb randomly and then LoginEventProcess must login on the Alfresco Share application with these different randomly picked users.

I have also provided session delay according the formula mentioned in its doc :-

bmshare.test.sessionDelay=10000
bmshare.test.user.session.max=280000
bmshare.test.user.session.min=200000
bmshare.test.user.think.max=38000
bmshare.test.user.think.min=2000

To calculate number of concurrent sessions:
S:  mean user session time
N:  number of concurrent users
T:  time between session starts (bmshare.test.sessionDelay)

N = S/T

In the above example:
S = 240s
N = 60 (desired load)
T = S/N = 4s
So set 
bmshare.test.sessionDelay=4000

accordingly with my parameters. let me know if there is any solution of this. Share Scenario Parameters

share load parameters

WebDrone Parameters

MongoDb Parameters

1

There are 1 answers

0
Derek Hulley On

The "Number of Users" is simply a check to ensure that you have at least that number of users in the MongoDB users collection.

The user collection from the Mirrors section determines which users will be used to log in. Just like the other types of tests, it is a prerequisite that you sign up a bunch of users - however many you need - and then the Share test will randomly pick from the collection of created users. If you have created 1000 users, you can be pretty sure that the Share test will be using different users for each session.

The users in the Share test will create sites but they will not be granting access to those sites to other users. So if you need users to work against the same site, you need to first create your users (say 100) and then use the dataload test to create your site(s) and specify the number of site members as 100. All the users will thus be members of all the sites.

Then, in the Share test, set the "Select Site Create" weighting to zero, so that no sites will be created by the Share users. All users logging in will have access to the same sites. If you only create one site, they will all be hitting the same site.

So, to summarize:

  1. Use the Enterprise Signup test to create N users
  2. Use the dataload to create S sites with N users per site
  3. Disable site creation in the Share test
  4. Ignore the number of users check
  5. Configure you user concurrency