stack table's uuid primary key wastes resources in other tables

[Clint Byrum]

The stack table should just have an integer primary key. Every event, for instance, has to have 36 bytes of UUID instead of a 4 byte integer pointer back to the stack. The same is true of every resource record and others as well.

The UUID should just be a secondary unique index.

[Steve Baker (steve-stevebaker)]

Setting this to High priority

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/106054

[Sirushti Murugesan (sirushtim)]

So it seems like this is a necessity now for the current convergence architecture because of how the SyncPoint database schema is modeled. Reassigning this to myself because no activity for ~1 year.

[Steve Baker (steve-stevebaker)]

Setting this back to High, since it is needed for convergence

[Zane Bitter (zaneb)]

I guess this wasn't really _needed_ for convergence, because we now have convergence but we don't have this :/

Setting it back to Medium.

[OpenStack Infra (hudson-openstack)]

Change abandoned by Steve Baker (<email address hidden>) on branch: master
Review: https://review.openstack.org/106054
Reason: This is a very old review which has unaddressed review feedback, failing tests, or merge conflicts.

For this reason it is being abandoned to make the review queue more maintainable. It can be restored and refreshed at any time.