2018-04-03 12:08:52 |
wangxiyuan |
description |
Backing to the patch https://review.openstack.org/#/c/399684/ when the domain_id is added for Idp, I assume that the domain_id is designed unique.
The domain_id in Idp model is unique: https://github.com/openstack/keystone/blob/master/keystone/federation/backends/sql.py#L58
But the db migration is not:
https://github.com/openstack/keystone/blob/master/keystone/common/sql/expand_repo/versions/012_expand_add_domain_id_to_idp.py#L55-L56
They are not in consistence. Since I don't know the origin purpose for the change, I'm not very sure which is correct but I prefer the first one.
What's more, our unit test use resource models to init db so that it's unique in our unit test sqlite db. It's not in consistence as well with the real usage. |
Backing to the patch https://review.openstack.org/#/c/399684/ when the domain_id is added for Idp, I assume that the domain_id is designed for unique?
The domain_id in Idp model is unique: https://github.com/openstack/keystone/blob/master/keystone/federation/backends/sql.py#L58
But the db migration is not:
https://github.com/openstack/keystone/blob/master/keystone/common/sql/expand_repo/versions/012_expand_add_domain_id_to_idp.py#L55-L56
They are not in consistence. Since I don't know the origin purpose for the change, I'm not very sure which is correct but I prefer the first one.
What's more, our unit test use resource models to init db so that it's unique in our unit test sqlite db. It's not in consistence as well with the real usage. |
|