Intermittent failure in CorruptConfigWithClassRuleIT

[jd0-sag]

The 1.0.11-beta2 release build had a test failure in CorruptConfigWithClassRuleIT. Given that the changes in this version modified how very early server failure is handled, I presume this was related.

Locally, this problem hasn't manifested and, if anything, the changes should have made the start-up more deterministic than it previously was.

Some investigation will have to be done to find this problem and determine why forcing the failure earlier didn't make this test more reliable.

[jd0-sag]

Tried running the tests in a loop, locally, on the weekend: no failures after 65 runs. I even tried adding some strategy sleeps to introduce perturbation of the relative operation ordering in the hopes to observe a timing hole but still saw no issue (the waits are also now in places where it wasn't clear how these could cause problems).

Today's release build didn't have a problem.

Given that it wasn't clear how the failure was triggered (the output looked correct), it might turn out that this was an environmental red herring. Leaving this open in case we see the problem again, in the future, but this is a puzzling one.