Comment 77 for bug 1253896

A five minute timeout seems excessive, but we may need that for reducing the gate failures until we can figure out why the servers take so long to respond.

Here is another occurrence that happened when there were only a couple of other jobs in the gate:
http://logs.openstack.org/17/143017/5/check/check-tempest-dsvm-neutron-full/003dcbe/console.html