Image API tests all error after recent Devstack change
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Glance |
Fix Released
|
Critical
|
Jay Pipes | ||
tempest |
Fix Released
|
Undecided
|
Jay Pipes |
Bug Description
After a recent devstack change that removed the v1/ from the image endpoint, all Tempest image API tests are failing:
https:/
Look like this:
16:55:40 =======
16:55:40 ERROR: Negative tests for invalid data supplied to POST /images
16:55:40 -------
16:55:40 Traceback (most recent call last):
16:55:40 File "/opt/stack/
16:55:40 results = self.client.
16:55:40 File "/opt/stack/
16:55:40 res = self.do_
16:55:40 File "/opt/stack/
16:55:40 return func(self, *args, **kwargs)
16:55:40 File "/opt/stack/
16:55:40 headers=headers)
16:55:40 File "/opt/stack/
16:55:40 return func(self, method, url, body, headers)
16:55:40 File "/opt/stack/
16:55:40 raise exception.
16:55:40 MultipleChoices: The request returned a 302 Multiple Choices. This generally means that you have not included a version indicator in a request URI.
16:55:40
16:55:40 The body of response returned:
16:55:40 {"versions": [{"status": "EXPERIMENTAL", "id": "v2", "links": [{"href": "http://
Apparently, we now need to update the tempest.
Changed in glance: | |
milestone: | none → folsom-1 |
status: | Fix Committed → Fix Released |
Changed in tempest: | |
status: | Fix Committed → Fix Released |
Changed in glance: | |
milestone: | folsom-1 → 2012.2 |
The root cause of this issue is that glance. common. client overwrites the management_url with the auth plugin's returned URL, and then calls glance. common. client. Client. configure_ url() with that management URL, *even if a URL was already supplied*.