Comment 4 for bug 1216851

Some investigation was performed to rule out issues with the machines where these tests run. One machine where a failure had been witnessed for the nova py27 tests in the gate pipeline (timing out both mysql and postgres migrations) showed nothing out of the ordinary with the system. When running a subsequent instance of the same job, it succeeded and system load was less than the number of vCPUs on that VM. The database logs are littered with all manner of critical/fatal errors, but I have no idea which ones are considered "normal" for these tests and which could indicate the test accidentally doing something wrong (rather than purposefully).