phoenix tests failing due to publishing hanging while using DataCite

[pdurbin]

At standup I mentioned that phoenix tests are failing right now which always makes me nervous if we're close to a release and 4.10 is imminent. The current commit is 0b625dd but the new failures (new since the single "testPrivateUrl" failure in #5400 I mean) appear to be due to publishing hanging while using DataCite rather than anything in the code.

Pull request #5399 already contains a line to switch phoenix over to the new "FAKE" PID provider so it can be a fix for this issue:

For anyone unfamiliar with the phoenix server, please see http://guides.dataverse.org/en/4.9.4/developers/testing.html#the-phoenix-server

To give a sense of the failures, here are screenshots from

https://build.hmdc.harvard.edu:8443/job/phoenix.dataverse.org-apitest-develop/311/

Part of the challenge for me is that when I look at https://status.datacite.org I don't know which of the currently 4 "Major Outage" indicators I should expect to affect the phoenix server, if any. Here's a screenshot:

I'd like to point out that while we could switch phoenix back to EZID, we switched it away from EZID for a reason in #4725. We were being told phoenix was DOS'ing EZID. For all I know, the phoenix IP address is still blocked by EZID.

The bottom line is that merging pull request #5399 should get the tests passing again on phoenix.

[pdurbin]

Now that pull request #5399 (and #5407) has been merged, all API tests are passing on phoenix again (build 312 in the screenshot below).