HDDS-10080. Let junit.sh fail if no tests were matched in repeated run

[adoroszlai]

What changes were proposed in this pull request?

Repeated tests in junit.sh should fail fast if no tests are executed in the first iteration. The goal is to indicate if the wrong test filter was used (e.g. abstract test class).

https://issues.apache.org/jira/browse/HDDS-10080

How was this patch tested?

Local tests

Non-existent class:

$ ITERATIONS=3 OZONE_REPO_CACHED=true \
  ./hadoop-ozone/dev-support/checks/junit.sh \
  -am -pl :hdds-config -Dtest='NoSuchTest'
...
[INFO] BUILD SUCCESS
...
Iteration 1 exit code: 1

$ cat target/unit/iteration1/summary.txt 
No tests were run

Non-existent method in existing class:

$ ITERATIONS=3 OZONE_REPO_CACHED=true \
  ./hadoop-ozone/dev-support/checks/junit.sh \
  -am -pl :hdds-config -Dtest='TestConfigFileAppender#noSuchTest'
...
[INFO] Tests run: 0, Failures: 0, Errors: 0, Skipped: 0
...
[INFO] BUILD SUCCESS
...
Iteration 1 exit code: 1

$ cat target/unit/iteration1/summary.txt 
No tests were run

Valid test class:

$ ITERATIONS=3 OZONE_REPO_CACHED=true \
  ./hadoop-ozone/dev-support/checks/junit.sh \
  -am -pl :hdds-config -Dtest='TestConfigFileAppender'
...

$ cat target/unit/summary.txt                                                                                                                   
Iteration 1 exit code: 0
Iteration 2 exit code: 0
Iteration 3 exit code: 0

flaky-test-check

Used flaky-test-check from HDDS-10079.

Non-existent test method:
https://github.com/adoroszlai/ozone/actions/runs/7584500146/job/20658460541

Abstract test class:
https://github.com/adoroszlai/ozone/actions/runs/7584281925/job/20657631720

Specific method in abstract test class:
https://github.com/adoroszlai/ozone/actions/runs/7584288405/job/20657650841

Valid test class and method:
https://github.com/adoroszlai/ozone/actions/runs/7584274733/job/20657610659

Regular CI

https://github.com/adoroszlai/ozone/actions/runs/7584240425

[sadanand48]

Thanks @adoroszlai for the fix, LGTM

[adoroszlai]

Thanks @sadanand48 for the review.