Pulp

The following Pulp Smash test targets this issue: http://pulp-smash.readthedocs.io/en/latest/api/pulp_smash.tests.rpm.api_v2.test_export.html#pulp_smash.tests.rpm.api_v2.test_export.ExportDistributorTestCase

This pull request updates the test case to accommodate this bug: https://github.com/PulpQE/pulp-smash/pull/259

The bug report above states that "Pulp should return an HTTP 3XX response when the second publication request is received." This is incorrect. The sentence should say "HTTP 4XX."

export_distributor is responsible for publishing content into a directory meanwhile yum_distributor is responsible to publish content via http/https, so that is logical that export_distributor will not work for content publish via http.

So i followed your steps and by having just export_distributor i was able to successfully export content into a specific dir, or into an ISO which is served via http.
When i tried to make a regular publish, pulp obviously complained that it misses the distributor that is responsible for that work.

$ pulp-admin rpm repo publish run --repo-id 123 

+----------------------------------------------------------------------+
                      Publishing Repository [123]
+----------------------------------------------------------------------+

This command may be exited via ctrl+c without affecting the request.

Task Failed

Missing resource(s): repo_id=123, distributor_id=yum_distributor

With that said, i think we should close it as not a bug as by design we have specific distributor responsible for specific things. Or maybe i don't follow something in the bug description.

After some chat and details clarification with Jeremy, we have an issue that appeared due to implementation of no-op publish feature. I will update the description of the issue accordingly.

I was mistaken when I stated that this bug affects Pulp 2.8.3 and master (2.9). That's untrue. This bug only affects Pulp 2.9.