HDDS-10123. InaccessibleObjectException in tests using ChecksumByteBufferImpl with Java 17

[adoroszlai]

What changes were proposed in this pull request?

Several tests fail when run with Java 17 due to:

java.lang.ExceptionInInitializerError
	at org.apache.hadoop.ozone.common.TestChecksumImplsComputeSameValues.testCRC32CImplsMatch(TestChecksumImplsComputeSameValues.java:66)
	at java.base/java.lang.reflect.Method.invoke(Method.java:568)
	at java.base/java.util.ArrayList.forEach(ArrayList.java:1511)
	at java.base/java.util.ArrayList.forEach(ArrayList.java:1511)
Caused by: java.lang.reflect.InaccessibleObjectException: Unable to make field boolean java.nio.ByteBuffer.isReadOnly accessible: module java.base does not "opens java.nio" to unnamed module @75eeccf5
	at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:354)
	at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:297)
	at java.base/java.lang.reflect.Field.checkCanSetAccessible(Field.java:178)
	at java.base/java.lang.reflect.Field.setAccessible(Field.java:172)
	at org.apache.hadoop.ozone.common.ChecksumByteBufferImpl.<clinit>(ChecksumByteBufferImpl.java:44)
	... 4 more

ozone command does open some packages for ChecksumByteBufferImpl to work (see #3759). This change applies the same logic (extracted to ozone-functions.sh) in junit.sh to Surefire argLine.

https://issues.apache.org/jira/browse/HDDS-10123

How was this patch tested?

Ran tests that failed previously with Java 17.

Example:

$ JAVA_HOME=/usr/lib/jvm/java-17-openjdk-amd64 ./hadoop-ozone/dev-support/checks/junit.sh -am -pl :hdds-common -Dtest='TestChecksumImplsComputeSameValues'
...
Java version: 17.0.9
...
Tests run: 2, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 1.206 s -- in org.apache.hadoop.ozone.common.TestChecksumImplsComputeSameValues

CI (also verifies that refactoring bin/ozone did not cause regression):
https://github.com/adoroszlai/ozone/actions/runs/7505840912

[sodonnel]

Is it intended for this branch to run for Java 17 as well as Java 9?

I think it will result in two --add-opens switches in the args:

--add-opens java.base/java.lang=ALL-UNNAMED"
--add-opens java.base/java.nio=ALL-UNNAMED"

I think this is OK, but just checking.

[adoroszlai]

Yes. The first part (specific to Java 17) was added for HDDS-8533. Ozone Cluster components (SCM, OM, DNs) fails to start with JDK 17. The second part got added later for HDDS-7228. ChecksumByteBufferImpl.update() is expensive. I just moved these from ozone to ozone-functions.sh.

[sodonnel]

Cool - sounds good.

[sodonnel]

LGTM - I just had one clarification on a part of the change.

[adoroszlai]

Thanks @sodonnel for the review.