"LR" OSGi performance Test Example

Actually, we're two.
Ø determine the maximum number of concurrent users of test login;
Average transaction response time (two queries)

How do you get this task to start a test job?

First, windowsresources settings (in fact, do not monitor the setting of the line I feel)
1. Add a Device
2. Windows Server 2008 Select Platform as Win2003 (this has been tried several to get the right)
3. Server user name and password
4. Select the Monitoring project


Second, set the Runtime-settings (before each run to check)
1. Change the timeout to 200
2. Set think time selection ignore
3. Set log selection extended log below 3 According to the circumstances of the selection (later study of the log feel set not set also does not matter)

Third, concurrent testing

The collection point is to be outside the transaction and before the transaction.

Recording steps:

1. Open the login page (LR opens automatically)

2. Enter User name password

3. Inserting a collection point

4. Insert Start transaction

5. Click the login button

6. Login Page Loading complete

7. Close the page

8. Insert End Transaction

If you are very familiar with the role of each piece of script code, you can add: Lr_rendezvous collection point function in the script.

Iv. Setting up concurrent scenarios

Gradually increase by 100
Start Vusers:start Vusers simultaneously
Duration:run until completion

Reference Merge graph: Number of runs-clicks per second (right-click Merge graphs)

Number of operations – average transaction corresponding time

V. Load recording

Recording steps:

1. Open the login page (LR opens automatically)

2. Enter User name password

3. Insert Start transaction

4. Click the login button

5. Login Page Loading complete

6. Insert End Transaction

7. Input data

8. Insert Start transaction

9, test function point

10. Insert End Transaction

11. Repeat Operation 7-11

12. Close the page

Vi. Setting the Load scenario

Up to 20 users per second, depending on the number and time of business modifications

Start Vusers:start vusers, 2 every 00:00:15 (HH:MM:SS)
Duration:run for 00:30:00 (HH:MM:SS)
Stop vusers:stop all vusers:5 every 00:00:30 (HH:MM:SS)

Reference Merge graph: Hits per second-throughput

Refer to a lot of examples out of 3 scenarios results are almost one is 1 seconds plus 1 users, one is 1 seconds plus 10 users, specifically how to set up here still do not understand

Vii. parameterization

1, need to parameterize the place right-click to select ReplaceWith parameter

2. Enter a name

3. Select Edit Withnotepad

4. Setting rules

There are three options in the Select Next row: Sequential (sequential read), random (randomly read), unique (unique)

There are also three options in updatethe value: Each occurrence (per encounter), each iteration (per iteration), Once (always)


Concurrent scenes run for passed state, load scene for stopped state (tangled I long time to no entanglements, load tangled look online said this is right)

My report is as follows
Ø use 100, 200, 500 users concurrent all through, using 1000 users and through 600 users;
Ø Response time: simple operation <=5s (login, user management query, log management query)

Conclusion: The test verifies that the number of concurrent users of the system login module is about 600, the user query in the number of users is 60 o'clock response time operation less than 3 seconds, log query in the number of logs is 180 when the response time operation is less than 3 seconds. (I feel the amount of data is too small)

"LR" OSGi performance Test Example