)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"7317cab254529f087f2197e048984074c8816bea","unresolved":false,"context_lines":[{"line_number":16,"context_line":"This patch adds a functional test that reproduce the bug."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"[1] https://review.opendev.org/#/c/511358/"},{"line_number":19,"context_line":"[2] https://review.opendev.org/#/c/606161/"},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"Related-Bug: #1852207"},{"line_number":22,"context_line":""}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"3fa7e38b_2d7040a4","line":19,"updated":"2019-11-18 18:45:13.000000000","message":"What\u0027s odd is that the instance.availability_zone shouldn\u0027t really have anything to do with the reschedule since when we reschedule and conductor calls the scheduler for another host, it should be filtering on the user-requested availability zone from the RequestSpec, not what is set on the instance.\n\nI guess the issue is really that the RequestSpec is not passed from compute to conductor on reschedule (that was added in Queens Ie5233bd481013413f12e55201588d37a9688ae78) and without that we build the request spec from the instance:\n\nhttps://github.com/openstack/nova/blob/stable/pike/nova/conductor/manager.py#L547\n\nhttps://github.com/openstack/nova/blob/6e2ec5c5a397b508cddb646e865b6f4d2132f1d6/nova/objects/request_spec.py#L156\n\nSo that\u0027s really the issue and why nulling out the instance.AZ in compute affects the RequestSpec in conductor during reschedule.\n\nSo yes https://review.opendev.org/#/c/606161/ helps, but so would Ie5233bd481013413f12e55201588d37a9688ae78.\n\nNote that with either of those Queens changes, if RPC API versions are pinned we\u0027d still have the same issue because conductor wouldn\u0027t get the request_spec or host_list parameters.","commit_id":"eefe3d8e134887348fea637a94c7ed561dd229dc"}],"nova/tests/functional/regressions/test_bug_1852207.py":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"7317cab254529f087f2197e048984074c8816bea","unresolved":false,"context_lines":[{"line_number":96,"context_line":"            self.assertEqual("},{"line_number":97,"context_line":"                \u0027compute_not_in_my_az\u0027, server[\u0027OS-EXT-SRV-ATTR:host\u0027])"},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"            self.assertTrue(mock_instance_claim.called)"}],"source_content_type":"text/x-python","patch_set":1,"id":"3fa7e38b_5b02150e","line":99,"updated":"2019-11-18 18:45:13.000000000","message":"nit: mock_instance_claim.assert_called_once() might be a bit better","commit_id":"eefe3d8e134887348fea637a94c7ed561dd229dc"}]}