<div dir="ltr">Dear fellow Neutron'ers and QA'ers,<br>
<br>During the Atlanta Design Summit we have been assigned 20 minutes (<a href="http://junodesignsummit.sched.org/event/48ccd60090740ae80b4d1811b9a61303#.U12EsqbwBPq" target="_blank">http://junodesignsummit.sched.org/event/48ccd60090740ae80b4d1811b9a61303#.U12EsqbwBPq</a>)
to agree on the Tempest testing that will be developed for Neutron
during the Juno cycle. In order to make the most out of those 20
minutes, we want to start the conversation ahead of time, so, to the
extent possible, we concentrate on reaching agreement during the Atlanta
session. To get the conversation rolling, here's an initial list of
topics where we, as a community, need to reach consensus:<br>
<ul><li>Scenario testing. While during Icehouse we achieved a good level
of community engagement and coverage in API testing, scenarios have
received little attention, even though a few developers made great
contributions. During Juno, we want to significantly expand this effort,
along the following lines:</li><ul><li>We are looking for ideas for new scenarios from anyone and everyone (dev, qa, automantion,
manual, users, etc). There are no bad ideas. We need ideas, not
necessarily fully formed blueprints, though the latter would be even better. Don't let constraints (whitebox, multi-host, etc) to refrain you from proposing an idea. We will sort
through them later. Ideas are our initial gap right now. <br></li><li>Creation of blueprints for the agreed upon
scenarios, so potential contributors can volunteer to implement them and
progress tracking can be accomplished.</li><li>Creation of a "how to" or "primer" wiki page on how to implement Neutron scenario tests <br>
</li><li>Documentation of scenario tests. <span>While api
tests are to a great extent self explanatory, scenarios are more complex and it's not easy
for people other than the writers of a specific test to understand. We need to improve documentation.</span></li><ul><li><span>One solution might be to assign scenario tests owners to keep them up to date and well documented<br>
</span></li></ul></ul><li><span>API tests. The challenge in this area seems to be in:</span></li><ul><li><span>Closing the gaps that might haven been left open during Icehouse</span></li><li><span>Adding new tests needed as a consequence of changes and evolution of the Neutron API</span></li>
<li><span>Define an on going process to prevent api tests to become outdated or stale</span></li></ul><li><span>Nova Networking - Neutron parity sub-project. Are there
any specific needs in this sub-project that can be covered with Tempest
based testing?</span></li><li><span>Other Neutron sub-projects. Are
there specific needs of other Neutron sub-projects that can be covered
with Tempest based testing?</span></li></ul><p>This is a list of topics meant to start the conversation on this
subject. Please feel free to chime in, either in the mailing list or at
this etherpad page <a href="https://etherpad.openstack.org/p/TempestAndNeutronJuno" target="_blank">https://etherpad.openstack.org/p/TempestAndNeutronJuno</a><br>
</p><p>Thanks in advance for your input</p><p>Miguel Lavalle<br></p></div>