For the tempest restructure I moved out the scenario tests to a new directory. This means we change run order so they happen after the tempest/tests/volume directory, instead of before (they used to be in compute/servers and network). The results seem to be a very regular race in volume resource exhaustion. https://review.openstack.org/#/c/28624/ is the review, which was failed (though only on the quantum run) http://logs.openstack.org/28624/15/check/gate-tempest-devstack-vm-quantum/22583/logs/screen-c-sch.txt.gz - 2013-05-15 15:00:05 (you can see the scheduler issue) I'll be digging into this today, but any other thoughts around this would also be appreciated, especially if anyone else has noticed it. -Sean -- Sean Dague http://dague.net