Cloned from https://bugzilla.redhat.com/show_bug.cgi?id=1494688

Description of problem:

It is possible for duplicate entries to be placed in taxable_taxonomies table.

Example:

id   | taxonomy_id | taxable_id | taxable_type |         created_at         |         updated_at         
-------+-------------+------------+--------------+----------------------------+----------------------------
 17766 |          15 |         76 | Subnet       | 2017-08-10 09:05:41.651359 | 2017-08-10 09:05:41.651359
 17767 |          15 |         76 | Subnet       | 2017-08-10 09:05:41.687729 | 2017-08-10 09:05:41.687729

Notice the consecutive ids and virtually identical created_at timestamps

Version-Release number of selected component (if applicable): Satellite 6.2.11

How reproducible: No known reproducer - but seems like the point in time which the row is created is a good starting point. Possible that a unique constraint could be placed around (taxonomy_id, taxable_id).

Steps to Reproduce:
1.
2.
3.

Actual results: Duplicate entries as shown in the example

Expected results:

select taxable_type, taxable_id, taxonomy_id, COUNT() as Duplicates from taxable_taxonomies GROUP BY taxable_type, taxable_id,taxonomy_id HAVING COUNT() > 1;

This query should not ever return results.

Additional info: This manifests via APIs which deal with taxonomies such as Locations

curl -X GET -k -u admin https://localhost/api/v2/subnets/:subnet_id/locations"