Above is a collection of log files from the current tempest run. Within them should be everything you need to get to the bottom of a test failure. The screen-* logs will be your most valuable tools in this process. Use the timestamp of the failed test in the current tempest run.
If there is a compute test failure, especially a server not getting created correctly, or being in an unexpected state, the following is typically the most fruitful order to look at things:
If there is a volume failure in the test, the following is typically the most fruitful order to look at things:
This help file is part of the openstack-infra/system-config project, and can be found at modules/openstack_project/files/logs/help/tempest-logs.html . The file can be updated via the standard OpenStack Gerrit Review process.