Fix PEM certificate generation and mounting in OpenShift

[michalvavrik]

Summary

Fixes:

• PEM NPE in OCP: I have fixes in Quarkus QE Test Suite that requires FIPS-compatible PEM certificates. However right now the certificate library we use to generation returns path to the some-prefix-server-ca.crt even though it doesn't exist and so mounting in OpenShift fails over missing file.
• PEM cert and key mounting: private key and certificates are not mounted for us. This fixes that issue.
• scope of certificate config properties: I mentioned that when tests share same base class as Vert.x JWT tests do, all previously generated certs are mounted to OCP because the scope of properties set via io.quarkus.test.bootstrap.BaseService#withProperty(java.lang.String, java.lang.String) is all the tests of all test classes (AKA RestService is stored in a static field). This PR avoids mounting of certs from previous tests.

Here are tests quarkus-qe/quarkus-test-suite@774d6d0 I run both baremetal (in FIPS) and OpenShift tests with this change and they pass.

Please check the relevant options

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Dependency update
• Refactoring
• Breaking change (fix or feature that would cause existing functionality to change)
• This change requires a documentation update
• This change requires execution against OCP (use run tests phrase in comment)

Checklist:

• Example scenarios has been updated / added
• Methods and classes used in PR scenarios are meaningful
• Commits are well encapsulated and follow the best practices

[michalvavrik]

run tests

[michalvavrik]

@mjurc Run OpenShift Tests in JVM is notoriously unstable these days, I personally run it more than 10 times this week in various PRs (see the job history) and it often fails over Jenkins job failure:

FATAL: command execution failed
java.io.EOFException
	at java.base/java.io.ObjectInputStream$PeekInputStream.readFully(ObjectInputStream.java:2911)

Not sure what is going on there, but considering the native job succeeds I think this PR is alright. Even though I retriggered it manually, I wouldn't wait.

[mjurc]

Thanks for the update.