Better handle 404 from Pulp and 410 GONE from Candlepin when removing objects
|Target version:||Team Brad - Backlog|
|Velocity based estimate||-|
|Release||Katello Backlog||Release relationship||Auto|
Whenever we try to delete an object in Pulp and Candlepin, there are situations where there is a pre-existing data mismatch because a user skipped a paused step or Force Unlocked a stuck task.
There are also situations where you try to delete or disable objects and the actual delete fails because it tries to call DELETE on the remote object via REST and gets a 404.
For cases with a delete hits a 404 trying to remove an object in Pulp/Candlepin, we should continue with the operation and not fail. If the object is gone in the remote API, then just finish out the transaction and delete the object
An example test case to induce this failure state:
1) Enable a repo in the Red Hat Repositories
2) use pulp-admin to delete said repo
3) Attempt to disable repo in Red Hat Repositories page and get a 404
You can run into this when trying to delete higher level containers such as an Organization delete. If any contained remote object in Candlepin or Pulp is missing, the Org delete will fail on 404 errors and the user is forced to hand-skip each step in the transaction until all missing object deletes are skipped.
We need to keep processing the removal and not fail.