What do you do if you try to run a test, but it fails to start? You may have been met with a Failed Report Summary showing this:
Another variation of this behavior may instead result in an empty report being generated:
Alternatively, you may have observed that your test hangs indefinitely at the starting stage, generating the above message only after you manually terminate it.
When this happens, the first thing to do is to click the Logs tab to review your Logs Report. Download the artifacts.zip and review the logs included in it.
Note: Terminating a test will prevent it from generating an artifacts.zip. You may need to check for all the possible causes below in the absence of a log to point you to which one is likely in play.
The most common causes of a startup hang or failure fall under these categories:
The #1 most common cause for a startup hang or failure is a missing file, such as a missing CSV file that your test script references, for example. The tell-tale sign that this is the problem at play is when you find this in one of your log files (usually the bzt.log and/or -- when running a JMeter test -- the jmeter.log):
2019-05-01 21:21:40,919 ERROR o.a.j.t.JMeterThread: Test failed! java.lang.IllegalArgumentException: File your_file_name.csv must exist and be readable
This means one of the following:
- You may not have uploaded all files required to run your test script in BlazeMeter. If your test script requires any additional files (CSVs, JARs, etc.), then refer to our guide on Uploading Files & Shared Folders.
- You may have uploaded a required file, but your test script still refers to it using a local file path. (For example, a CSV Data Set Config element may still refer to a CSV file using a local path.) This is likewise explained in our guide on Uploading Files & Shared Folders.
- You may have uploaded a required file, but the filename referenced in your script is in a different case than the filename of the one uploaded. BlazeMeter engines run on Linux, which treats filenames as case-sensitive, so Your_File_Name.csv is a different file from your_file_name.csv or Your_File_name.csv. Make sure your filename references are consistent with the case used in your uploaded filename.
- You may have accidentally designated the wrong file as your main test file. In the following example, we have a Taurus YAML file pointing to a Gatling script file, and the YAML is correctly designated as the main file:
However, if you accidentally click the other file, the arrow will move to it, changing it to the main file:
The second scenario will fail to start since Blazemeter requires a YAML in order to execute the Gatling script, yet in this configuration, Blazemeter will ignore the YAML file entirely. To fix it, just click the YAML again to reinstate it as the main test file.
- You may have tried to run a test that requires more engines than your plan allows for. For example, if you are on a Basic Plan which only allows for two engines (load generators), but attempt to start a test configured to utilize four engines, the test may hang on startup since it is trying to spin up two additional engines that are not available under the terms of the subscription plan. Note: If you have enabled End User Experience Monitoring, please be aware that that feature requires running on an additional engine of its own. For example, a one-engine test with EUX Monitoring enabled requires two engines in total.
If you're executing your test via a Private Location (aka on-premise location or OPL) instead of a cloud engine, then the problem may not be with your test, but with your agent.
- The most common problems with agents include connection issues or the agent running out of disk space, both of which can be addressed by ensuring all System Requirements are met and testing for any connection issues between the agent and BlazeMeter.
- Make sure to review your agent's logs and verify all image tags appears as they should.