Fix the race condition in e2e test.

[bashir2]

Based on this report:

There is a race condition between the BULK_EXPORT test and the controller test. Based on the dep graph and FHIR-server parameters, I think both of these tests are querying http://hapi-server:8080; but the problem is that in the controller test we are testing the incremental pipeline functionality as well. So depending on whether it runs before or after BULK_EXPORT test we may get extra resources in BULK_EXPORT test, which can fail the whole e2e test.