A test job in ACCELQ may include multiple test cases - either part of a Scenario with multiple data driven test cases or as a part of a test suite which is a collection of Scenarios. You get an indication of number of test cases in the Run modal when you are kicking off a test.
Each test case in ACCELQ runs as an independent flow, and it is possible to enable parallel execution of test cases to speed up run time. ACCELQ takes care of collating the results from multiple parallel threads, and present a unified report for the entire test job.
Step 1: Configure parallel execution for an agent
When you startup an agent, you can configure the number of parallel test cases in can accommodate. This is provided as part of the agent.properties file under your Local Agent folder. If you make a change to this property, be sure to restart the agent.
Step 2: Select parallel execution for a test job
Select the option to use Parallel Execution when you run a job. Depending on the agent you selected to run this test against, parallel instance count is allowed to be configured. You can set the parallel count for the current Run, up to a maximum threshold specified in the Agent startup.
Important Considerations in Parallel Execution
- Since each test case will open a new browser instance, ensure that you allocate sufficient system resources (memory, CPU) on the Agent machine to execute the intended parallel count.
- If the application-under-test is particularly sensitive to mouse movements (hovers etc.) or keyboard inputs, you may encounter conflicts while working against multiple concurrent browsers. Tune the parallel count accordingly.
- When you are running a Static Test Suite, parallel execution does not guarantee the order of execution of tests, since the test cases are picked up as and when a thread gets freed up.
Comments
1 comment
Hi.. Can I get more detail on below?
When you are running a Static Test Suite, parallel execution does not guarantee the order of execution of tests, since the test cases are picked up as and when a thread gets freed up.
How can we ensure the order of execution in suite (static/filterbased) when run in parallel?
Please sign in to leave a comment.