test(db): use h2:file db mode by default

[andrewazores]

Fixes #1094
Depends on #1083

[github-actions]

This PR/issue depends on:

• cryostatio/cryostat#1083
  By Dependent Issues (🤖). Happy coding!

[andrewazores]

Along with setting the h2:file mode to the default in smoketest.sh and the pom.xml itest setup, I also made a few minor changes to some itests themselves to disable credentials handling in the AutoRulesIT, at least temporarily. I'm not sure exactly why but that test is flaky with the interaction between credentials, target discovery, and rule activation, though in practice with manual testing I can never reproduce the failure.

I also removed some detach calls from within DAOs. The entities should be fine detached, but removing these calls also seemed to improve the test reliability somewhat.

There's more work to be done around the DAOs and how we use Hibernate/JPA in general to ensure we're handling data consistency properly and not leaving resources open, but that's outside the scope of this PR.

[maxcao13]

Along with setting the h2:file mode to the default in smoketest.sh and the pom.xml itest setup, I also made a few minor changes to some itests themselves to disable credentials handling in the AutoRulesIT, at least temporarily. I'm not sure exactly why but that test is flaky with the interaction between credentials, target discovery, and rule activation, though in practice with manual testing I can never reproduce the failure.

Is there any useful logs of these runs that I may be able to look at?

[andrewazores]

I didn't bother preserving any of the ones I observed locally, but this CI run looks like it's the same thing:

https://github.com/cryostatio/cryostat/actions/runs/3198632408/jobs/5223422453

[maxcao13]

Hmm, that's weird - not too sure as of now without diving too deep, but anyways this PR looks good!

[andrewazores]

Sorry, I just went back and checked on the JPA entity detachment thing and found that the change was no longer making any difference. It may have just seemed that way by coincidence yesterday when I was testing it since it was only affecting the flaky test(s) to begin with.

[andrewazores]

I also included a change to the repeated-integration-testing.bash script to reduce the JDP polling times and not require multiple verifications of the count. Those default settings are overly cautious and cause tests to run a fair bit slower than really necessary, so using that development re-testing script to experiment with speeding JDP things up could be interesting.

[maxcao13]

That makes dev testing a lot faster, right looks good!