make AggregateAPI work in cells

[Matthew Sherborne]

Make nova.compute.AggregateAPI work in child cells:

    def create_aggregate(self, context, aggregate_name, availability_zone):
    def get_aggregate(self, context, aggregate_id):
    def get_aggregate_list(self, context):
    def update_aggregate(self, context, aggregate_id, values):
    def update_aggregate_metadata(self, context, aggregate_id, metadata):
    def delete_aggregate(self, context, aggregate_id):
    def add_host_to_aggregate(self, context, aggregate_id, host_name):
    def remove_host_from_aggregate(self, context, aggregate_id, host_name):

Should all work through cells.

[Matthew Sherborne (msherborne+openstack)]

In progress: https://review.openstack.org/#/c/25813/

Todo list (* means to do) (+ means done):
********************************************************************************
Host aggregate working in cells
********************************************************************************
 * compute.cells_api.AggregateAPI
     + create_aggregate(self, ctxt, aggregate_name, availability_zone):
     + get_aggregate(self, ctxt, aggregate_id):
     + get_aggregate_list(self, ctxt, cell=None):
     + update_aggregate(self, ctxt, aggregate_id, values):
     + update_aggregate_metadata(self, ctxt, aggregate_id, metadata):
     + delete_aggregate(self, ctxt, aggregate_id):
     + add_host_to_aggregate(self, ctxt, aggregate_id, host_name):
     + remove_host_from_aggregate(self, ctxt, aggregate_id, host_name):
 * TEST - compute.cells_api.AggregateAPI
     * create_aggregate(self, ctxt, aggregate_name, availability_zone):
     * get_aggregate(self, ctxt, aggregate_id):
     * get_aggregate_list(self, ctxt, cell=None):
     * update_aggregate(self, ctxt, aggregate_id, values):
     * update_aggregate_metadata(self, ctxt, aggregate_id, metadata):
     * delete_aggregate(self, ctxt, aggregate_id):
     * add_host_to_aggregate(self, ctxt, aggregate_id, host_name):
     * remove_host_from_aggregate(self, ctxt, aggregate_id, host_name):
 * nova.cells.manager
     + create_aggregate(self, ctxt, aggregate_name, availability_zone):
     * get_aggregate(self, ctxt, aggregate_id):
     * get_aggregate_list(self, ctxt, cell=None):
     * update_aggregate(self, ctxt, aggregate_id, values):
     * update_aggregate_metadata(self, ctxt, aggregate_id, metadata):
     * delete_aggregate(self, ctxt, aggregate_id):
     * add_host_to_aggregate(self, ctxt, aggregate_id, host_name):
     * remove_host_from_aggregate(self, ctxt, aggregate_id, host_name):
 * TEST nova.cells.manager
     * create_aggregate(self, ctxt, aggregate_name, availability_zone):
     * get_aggregate(self, ctxt, aggregate_id):
     * get_aggregate_list(self, ctxt, cell=None):
     * update_aggregate(self, ctxt, aggregate_id, values):
     * update_aggregate_metadata(self, ctxt, aggregate_id, metadata):
     * delete_aggregate(self, ctxt, aggregate_id):
     * add_host_to_aggregate(self, ctxt, aggregate_id, host_name):
     * remove_host_from_aggregate(self, ctxt, aggregate_id, host_name):
 * nova.cells.messaging - Message Runner
     + create_aggregate(self, ctxt, aggregate_name, availability_zone):
     + get_aggregate(self, ctxt, aggregate_id):
     + get_aggregate_list(self, ctxt, cell=None):
     + update_aggregate(self, ctxt, aggregate_id, values):
     + update_aggregate_metadata(self, ctxt, aggregate_id, metadata):
     + delete_aggregate(self, ctxt, aggregate_id):
     + add_host_to_aggregate(self, ctxt, aggregate_id, host_name):
     + remove_host_from_aggregate(self, ctxt, aggregate_id, host_name):
 * TEST nova.cells.messaging - Message Runner
     * create_aggregate(self, ctxt, aggregate_name, availability_zone):
     * get_aggregate(self, ctxt, aggregate_id):
     * get_aggregate_list(self, ctxt, cell=None):
     * update_aggregate(self, ctxt, aggregate_id, values):
     * update_aggregate_m...

[Matt Riedemann (mriedem)]

The patch was abandoned but it sounds like rax still wants to do this in Juno, so marking as confirmed and removing assignee.

[Joe Gordon (jogo)]

now that we are planning on replacing the entire cells design with something else and don't want to spend time fixing something we are about to remove, closing this.

[Matt Riedemann (mriedem)]

Added here: https://etherpad.openstack.org/p/nova-cells-testing