Hi infra-ninjas, I missed this thread originally, not previously having been subscribed to os-infra ML.
So I just wanted to throw out onto the table a (possibly naïve) alternative way of looking at this issue. The core problem seems to be that the node allocation algorithm has a tendency to skew in favor of the classes under heavier demand. But I guess there's no particular reason, other than historical, that the f20 nodes aren't more widely used? For example the py27 unit tests for a selection of the projects, and/or some of the tempest jobs, could potentially be re-based onto f20. If we were to use f20 more widely in the gate (not to entirely supplant precise, more just to split the load more evenly) then would the problem observed tend to naturally resolve itself? Just a thought ... Cheers, Eoghan _______________________________________________ OpenStack-Infra mailing list [email protected] http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra
