Open Stack

On Tue, Jan 17, 2017 at 02:48:27PM +0000, Adriano Petrich wrote:
>    Mathieu,
>    Â  Â  That sounds exactly what we need. Do we run tempest or something on
>    those to validate it?

It doesn't currently run tempest, only some basic sanity tests (crud
operations where we create some resources for each service before the
upgrade, then check they are still there after the upgrade is completed).

In future we could probably add more validation, but we're constrained by
walltime of the job.

As Mathieu says this does provide at least partial coverage of deploying an
old undercloud version (e.g Newton) with a latest (trunk/ocata) undercloud
- hopefully we can adjust the upgrade test coverage to meet your needs and
  avoid the overhead of a completely new job.

Steve