No way to prevent duplicates in endpoints

[Vish Ishaya]

Keystone doesn't provide a way to ensure that we do not create duplicate endpoints. Imagine two concurrent workers attempting to create a keystone endpoint. There are no unique constraints for endpoints besides the id, but the id cannot be set externally, so there isn't an easy way to ensure only one gets created without a global lock

[David Stanek (dstanek)]

What would you expect the unique constraint to be based on? I could see url, url/enabled, and a few other combinations.

[Dolph Mathews (dolph)]

I would suggest pretty much everything except enabled: service_id + region_id + interface + url

Would that work, Vish?

[Raildo Mascena de Sousa Filho (raildo)]

I will follow the @dolph suggestion. ok?

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/134095

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/241748

[OpenStack Infra (hudson-openstack)]

Change abandoned by Henrique Truta (<email address hidden>) on branch: master
Review: https://review.openstack.org/241748
Reason: This change won't be necessary anymore, since the URL column won't be included in the Unique constraint

[Steve Martinelli (stevemar)]

bumping this from mitaka-rc1, as the patch that fixes it is WIP. this bug is not a blocker, we've lived with it for a few releases now.

[Steve Martinelli (stevemar)]

Automatically unassigning due to inactivity.

[Steve Martinelli (stevemar)]

This would break backwards compatibility.

[OpenStack Infra (hudson-openstack)]

Change abandoned by David Stanek (<email address hidden>) on branch: master
Review: https://review.openstack.org/134095
Reason: The bug is marked as won't fix