Open Stack

I feel that Ironic developers should not be concerned with data migrations
until we approach an RC. I do expect to have schema migrations to keep the
db consistent with code changes, but at this early stage with Ironic (when
much of the code is still being written and refactored, and the schema may
change significantly) I don't feel that there is any practical benefit to
supporting data migrations. I also plan to wait until we near an RC before
writing a nova -> ironic data migration tool. Doing so before our first RC
is just going to delay getting there.

Since everyone loves ASCII, here's an attempt at drawing what I'm talking
about w.r.t schema changes:

 nova_bm   --> ... --> ... --> "H" or "I" release
  (fork)                    (data migration script)
    |                               \
  ironic   --> 002 --> 003 --> ... --> RC1

I'd like to know how folks feel about this plan, and what other project
splits have done in this regard.


Regards,

--
Devananda
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20130522/70ecb874/attachment.html>