)]}'
{"nova/compute/api.py":[{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"cf441fa977489734bc92e539956c8267131687f6","unresolved":false,"context_lines":[{"line_number":2822,"context_line":"    def get_device_metadata(self, context, instance):"},{"line_number":2823,"context_line":"        \"\"\"Retrieve device metadata (including tags) for the given instance.\"\"\""},{"line_number":2824,"context_line":"        return self.compute_rpcapi.get_device_metadata(context,"},{"line_number":2825,"context_line":"                                                       instance\u003dinstance)"},{"line_number":2826,"context_line":""},{"line_number":2827,"context_line":"    @wrap_check_policy"},{"line_number":2828,"context_line":"    @check_instance_lock"}],"source_content_type":"text/x-python","patch_set":2,"id":"5a18252c_16bab123","line":2825,"updated":"2016-04-12 16:10:13.000000000","message":"So, this is mostly a concern further up the stack, but the solution reflects here I think. Making a blocking call from the API to the compute node for information that doesn\u0027t change very often is really something we should avoid. Especially since it\u0027s done in two places (here for the REST API and again in a later patch for the metadata API) I feel pretty strongly that this information needs to be written to the database and merely read here. It will be much faster and more efficient for both cases.","commit_id":"f743a294ef7bca5fd34e97f06216600a45c521b3"}]}