Stabilize test_scanjob_cancel

This test fails every few runs.

Sometimes it fails because job never starts - this is a known problem
with thread-based scan manager, and we can't do anything about this.

Sometimes it fails because we want scanjob to be running, but it
completed already. This seems to only happen for Ansible source. There's
nothing wrong with Ansible itself, but it seems that Ansible in our
environment is relatively small and tends to finish too quick. For now
let's ignore it as a source for this test.

camayoc/tests/qpc/cli/test_scanjobs.py

@@ -12,6 +12,7 @@
 import re
 
 import pytest
+from littletable import Table
 
 from camayoc.constants import QPC_OPTIONAL_PRODUCTS
 from camayoc.exceptions import NoMatchingDataDefinitionException
@@ -284,7 +285,12 @@ def test_scanjob_cancel(qpc_server_config, data_provider):
         3) Try to restart the scan by running ``qpc scan restart --id <id>``
     :expectedresults: The scan must be canceled and can't not be restarted.
     """
-    source = data_provider.sources.new_one({}, data_only=False)
+    # There's nothing fundamentally wrong with Ansible. In our environment,
+    # it finishes too quick, largely increasing a risk of test failing
+    # because job finished before we checked if it started.
+    source = data_provider.sources.new_one(
+        {"type": Table.not_in(("ansible", "openshift"))}, data_only=False
+    )
     scan_name = uuid4()
     scan_add_and_check({"name": scan_name, "sources": source.name})
     data_provider.mark_for_cleanup(Scan(name=scan_name))