To show the behavior of tests using data sets, consider the following scenarios:
A test has 15 virtual users, and a data set has two rows of data.
Scenario 1: At the end of data -> Start over
The first user would read the first row of data; the second user would read the second row. The third user starts over and reads the first row, and so forth. This cycle continues until all 15 users have run the test. The first row was read eight times; the second row seven times.
Scenario 2: At end of data -> Execute End step
The first user would read the first row of data, the second user would read the second row. The third user through the fifteenth user would start the test, and immediately jump to the End step.
A test has 100 virtual users, and the data set has 1500 rows of data. Tests are running concurrently.
Scenario 1: At the end of data -> Start over
When the 100 users start, they would read the first 100 rows of data. Depending on the staging document, as cycles end, the users start new runs of the test case and consume more rows of the data set. If all rows are consumed and the test run has not ended, it restarts with the first row until the test run ends.
Scenario 2: At end of data -> Execute End step
The test run ends after 1500 cycles.
A test has 10 virtual users, and the data set has 10,000 rows of data. The staging document specifies that the test will run for 2 minutes.
The 10 users start and read the first 10 rows of data and continue consuming rows of data from the 10,000 rows. How fast they run determines how far down the data set they go. At the two-minute mark, the test ends.
|
Copyright © 2014 CA Technologies.
All rights reserved.
|
|