Reduce duplicate data between unified limit tables
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Identity (keystone) |
Fix Released
|
Medium
|
wangxiyuan |
Bug Description
Unified limits consists of registered limits and limits, which are project-specific. Both types are isolated into separate database tables (keystone.
http://
There are some duplicated attributes between these two tables that are requirements for the API (e.g. a limit cannot be created without a corresponding limit being registered first). We have the opportunity to simplify the schema of the keystone.limit table by removing limit.service_id, limit.region_id, and limit.resource_name in favor of a single column called limit.registere
The unified limit API will use the FK relationship to populate service_id, region_id, and resource_name attributes of a limit from the corresponding registered_limit.
Possibly solutions for the migration were whiteboarded using the following etherpad:
https:/
Changed in keystone: | |
importance: | Undecided → Medium |
status: | New → Triaged |
description: | updated |
tags: | added: sql |
tags: | added: limits |
description: | updated |
Changed in keystone: | |
assignee: | nobody → wangxiyuan (wangxiyuan) |
Changed in keystone: | |
milestone: | none → rocky-rc1 |
Fix proposed to branch: master /review. openstack. org/577751
Review: https:/