make -e full run tempest scenarios serially
The scenario tests end up generating a large load on target
environments. In most cases this means the whole system is running at
an extremely unrealisticly high load, and many of our scenario
failures are probably traceable back to this.
This classifies many more scenario tests as slow so that during normal
full runs they are not run, they are still accessible for running with
the tox -e all flag.
Selection based on https://ethercalc.openstack.org/nu56u2wrfb2b
Depends-on: Ia303773cabffc25d910cd1c0950d6739592bade6
Change-Id: I35cf4a77ff61d04ed8be58acff3edf47b1ae36f7
diff --git a/tox.ini b/tox.ini
index d8d390e..ddda2ab 100644
--- a/tox.ini
+++ b/tox.ini
@@ -78,7 +78,8 @@
# See the testrepository bug: https://bugs.launchpad.net/testrepository/+bug/1208610
commands =
find . -type f -name "*.pyc" -delete
- tempest run --regex '(?!.*\[.*\bslow\b.*\])(^tempest\.(api|scenario))' {posargs}
+ tempest run --regex '(?!.*\[.*\bslow\b.*\])(^tempest\.api)' {posargs}
+ tempest run --combine --serial --regex '(?!.*\[.*\bslow\b.*\])(^tempest\.scenario)' {posargs}
[testenv:full-serial]
envdir = .tox/tempest