Open Stack

> That change works on the dataset. However I was referring to the
> db/object api (which I have no real knowledge of) that it should be able
> to get_by_uuid unmigrated instances and in my case I got the traceback
> given in that paste. It's possible I'm just using the API incorrectly.

You should be able to pull migrated or unmigrated instances, yes. The
tests for the flavor stuff do this (successfully) at length. The
traceback you have seems like a half-migrated instance, where there is a
non-null flavor column on the instance_extra record, which shouldn't be
possible. The line numbers also don't match master, so I'm not sure
exactly what you're running.

If you can track down where in your dataset that is hitting and maybe
figure out what is going on, we can surely address it, but the current
tests cover all the cases I can think of at the moment.

Any chance you're tripping over something that was a casualty of
previous failures here?

> Oh yes, we want that restriction, but a way around it for instances that
> may be stuck or just testing purposes could be useful.

Yeah, as long as we're clear about the potential for problems if they
run --force on a moving target...

--Dan