Description of problem:

This bug is even more relevant because the bootstrap.py script uses the /katello/api/capsules/<capsule_id> API endpoint to migrate hosts between capsules.

API requests to /katello/api/capsules/<some_capsule_id> fail if run by a non-admin user, even if said user has _all_ roles. As soon as we enable the admin flag for the user, the same API request works and returns information about the individual capsule.

In contrast, requests to /katello/api/capsules succeed without requiring the admin flag.

The information returned by the latter includes 100% of the information returned by the former, so the permission requirements for one of these two requests looks incorrect. I believe /katello/api/capsules/<capsule_id> should _not_ require the admin flag.


Version-Release number of selected component (if applicable):
Satellite 6.10 and earlier (tested on 6.9 and 6.8).


How reproducible:
100%

Steps to Reproduce:
1. Create a user on Satellite, give it all permissions but not _the_ admin flag.
2. Have an external capsule. In this example the external capsule ID is 2.
3. Query the API from Satellite itself at 2 points with the non-admin user:
    curl -k -u theuser:password https://localhost/katello/api/capsules/2
 and
    curl -k -u theuser:password https://localhost/katello/api/capsules/

Actual results:

The first curl command fails with:

{
  "error": {"message":"Resource smart_proxy not found by id '2'"}
}

The second command succeeds and returns all information about all capsules including capsule ID 2.


Expected results:

Both commands would succeed.



Additional info:

Location is not the issue here. These tests were performed on the lab and the user and capsule locations were made to match, or to differ. Location (mis)matching made no difference to the test outcomes, only the admin flag did.