)]}'
{"nova/tests/functional/regressions/test_bug_1713783.py":[{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"ff9f7dbc80b58103f16904ea566b3eb5be0d7536","unresolved":false,"context_lines":[{"line_number":124,"context_line":"        # restart the compute service to make sure it doesn\u0027t delete"},{"line_number":125,"context_line":"        # the instance based on the attempted evacuation"},{"line_number":126,"context_line":"        self.compute1.start()"},{"line_number":127,"context_line":"        self.api.get_server(server[\u0027id\u0027])"}],"source_content_type":"text/x-python","patch_set":1,"id":"9f6a8fd7_17d7f5ef","line":127,"updated":"2018-04-16 20:50:25.000000000","message":"Looking at the code in compute/manager\u0027s _destroy_evacuated_instances, it deletes only the local instance on the hypervisor (virt domain) along with the placement allocation records. IIUC, that won\u0027t necessarily cause a server GET to fail being that the instance record in the database is still around, even if the local instance (virt domain) was destroyed.","commit_id":"7a4117876a5e9e72b8956d16c0c03d48e7ab48b9"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"c9e42b10f9f6535872dd23c0f07f042d0e823c0e","unresolved":false,"context_lines":[{"line_number":124,"context_line":"        # restart the compute service to make sure it doesn\u0027t delete"},{"line_number":125,"context_line":"        # the instance based on the attempted evacuation"},{"line_number":126,"context_line":"        self.compute1.start()"},{"line_number":127,"context_line":"        self.api.get_server(server[\u0027id\u0027])"}],"source_content_type":"text/x-python","patch_set":1,"id":"9f6a8fd7_b9e95b81","line":127,"in_reply_to":"9f6a8fd7_17d7f5ef","updated":"2018-04-16 23:54:34.000000000","message":"Unfortunately, I don\u0027t think this will check what we intend, which is whether the local VM (virt domain) associated with the instance got destroyed during the restart. I think the instance record will be around either way.\n\nI think what you could do here instead of server GET is pull the migrations for the instance again and verify that the status is not \u0027completed\u0027 because the _destroy_evacuated_instances routine will set the migration status to \u0027completed\u0027 after it destroys the local instance and deletes the allocations from placement for the compute host.","commit_id":"7a4117876a5e9e72b8956d16c0c03d48e7ab48b9"}]}