On Monday, 17 April 2017 18:28:24 CEST Ben Nemec wrote: > On 04/17/2017 10:51 AM, Emilien Macchi wrote: > > We haven't got much feedback from TripleO core reviewers, who are > > usually more involved on this topic. I'll give a chance to let them > > talk because we take some actions based on the feedback brought in > > this discussion. > > I started to write a response last week and realized I didn't have a > coherent recommendation, but here are my semi-organized thoughts: > > The pingtest was created for two main reasons. First, it's fast. Less > than three minutes in most CI jobs. Second, it's simple. We've added a > bunch of stuff for resource cleanup and such, but in essence it's four > commands: glance image-create, neutron net-create, neutron > subnet-create, and heat stack-create. It would be hard to come up with > a useful test that is meaningfully simpler. > > Tempest isn't really either of those things. According to another > message in this thread it takes around 15 minutes to run just the smoke > tests. That's unacceptable for a lot of our CI jobs. It also tends to > require a lot more configuration in my experience.
I think that you are talking about the "full set of Tempest tests here". My point it is possible to have a test which has the exact same semantic as the current ping test, but written using tempest.lib and relying on the other tooling from Tempest to run it (tempest run/ostestr). With that in place, it would trivial to decide whether to run just that test or run other tests in other jobs (it would be a matter of a simple regexp) and it would simplify the code in other tools like tripleo-quickstart (no need to keep the preparation phase for two runners: validation and tempest). Ciao -- Luigi __________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
