Comment 3 for bug 1690781

Observation - once we hit this issue, all subsequent runs hit it (we don't restart or redeploy infra nodes between runs).
After I tried to reboot all infra nodes and run a test again, it passed this critical point.