As you suggested, I'll summarize here how we changed the db name for the integration tests in a simple form to begin with:
1) The DB name was hard-coded to "LSST" and created during the configuration procedure. The same name followed through into qservMeta and qservw_worker during their creation, as well as into the individual test cases.
2) The creation of these databses is removed from the configuration procedure in qserv.
3) In qserv_testdata, qservdataloader.py and benchmark.py have been updated to use an automated name, depending on the test case currently running.
4) The same name, of the form "qservTest_caseXX", is used to create the relevant database names during the tests (and not at config step), as well as table names in qservMeta and qservw_worker.
5) There are also some updates regarding the loading and deletion of these test cases so they do not interfere with each other and can be run individually in any sequence with "qserv-check-intergration.py --case=XX" or all at once with "qserv-test-integration.py"
I think this covers most of it, Fabrice Jammes may have some comments or corrections to what I've listed.