How to find the cause of (GlassFish) Platform TCK test failures

The https://ci.eclipse.org/jakartaee-tck/job/10/job/eftl-jakartaeetck-run-100 Jenkins job is used to run the EE 10 Platform TCK against GlassFish.

Even with only 437 failures shown on the https://ci.eclipse.org/jakartaee-tck/job/10/job/eftl-jakartaeetck-run-100/64, there are no test results copied to that jobs build artifacts due to there not being enough JNLP memory defined to hold the (large) test results. This wiki page will explain how to get the information that you need to understand more about why TCK tests failed.

1. Open the console link for the test run you are interested in. For example, https://ci.eclipse.org/jakartaee-tck/job/10/job/eftl-jakartaeetck-run-100/64/console and find the line that contains jakartaee-tck » master followed by started. (e.g. jakartaee-tck » master #1814 started.) and click on the hyperlink on that line (should open a job like https://ci.eclipse.org/jakartaee-tck/job/jakartaee-tck/job/master/1814).
2. Now click on the Build Artifacts link
3. You should see various (large) *results.tar.gz files that contain the test results + GlassFish Server.log), download each results file that you are interested in.
4. After extracting the contents, look at file root/run_cts.log to review the test failures if needed (perhaps note deployment file name for failed tests you are interested in).
5. Also review the root/vi/glassfish7/glassfish/domains/domain1/logs/server.log which contains the server log exceptions (look for the deployment name to key in on the tests your interested in).