Fix "Support pagination for 'list' API calls" backport

The change Ica7c8c953c11af4400b224ad942023f69a72ae02 was incorrectly
backported to stable/train, a merge conflict with the use of urllib from
six introduced a bug that triggers an exception when a list of octavia
resources exceeds the limit of the pagination:

module 'six.moves.urllib' has no attribute 'urlparse'

This commit fixes the invalid calls.

It also disables the lower-constraints job that is now failing on
train.

Note: this is a stable/train-only commit

Change-Id: I5523820e8607ed8e2fb75965427df33055673943

octaviaclient/api/v2/octavia.py

@@ -86,8 +86,8 @@ class OctaviaAPI(api.BaseAPI):
             links = response.get("{}_links".format(resource_key), [])
             for link in links:
                 if link.get('rel') == 'next':
-                    query_str = urllib.urlparse.urlparse(link['href']).query
-                    params = urllib.urlparse.parse_qs(query_str)
+                    query_str = urllib.parse.urlparse(link['href']).query
+                    params = urllib.parse.parse_qs(query_str)
                     break
             else:
                 break

zuul.d/projects.yaml

@@ -2,7 +2,6 @@
     templates:
       - check-requirements
       - openstack-cover-jobs
-      - openstack-lower-constraints-jobs
       - openstack-python-jobs
       - openstack-python3-train-jobs
       - openstackclient-plugin-jobs