fup: Increase service_down_time beyond INITIAL_REPORTING_DELAY in test

With service_down_time set to 5 seconds we could easily race the initial
report from the service that is controlled by
nova.servicegroup.api.INITIAL_REPORTING_DELAY which also defaults to 5
seconds.

This change simply increases service_down_time to 6 seconds to avoid
this race, ensuring disabled but alive computes are able to initially
report in and continue to be marked as up by the servicegroup API.

Closes-Bug: #1940741
Change-Id: I61fded6f513c18dbd0248ed54fc58a318dc05416

nova/tests/functional/regressions/test_bug_1938326.py

@@ -35,7 +35,7 @@ def _setup_compute_service(self):
         # We want the service to be marked down in a reasonable time while
         # ensuring we don't accidentally mark services as down prematurely
         self.flags(report_interval=1)
-        self.flags(service_down_time=5)
+        self.flags(service_down_time=6)
 
         # Use two compute services to make it easier to assert the call from
         # the dest to the src, we could also test this for same host resize.