Neutron Callback System

 用于core and service components之间的通信,传递resource的lifecycle events (e.g. before creation, before deletion, etc.)消息

,比如当一个Neutron resource与多个service (VPN, Firewall and Load Balancer) 关联,

需要Service的管理者来确定resource正确状态。

这样可以避免各个Service需要感知其他Service,在它们之间解耦和。

 它是一种进程内通信机制,如果是进程间则需要Messaging callbacks

如果是远程进程则需要 RPC API 。

比如service A, B, and C 都需要知道router creation,如果没有一个中介来采用消息的方式关联它们,

那么,A在router creation的时候就需要直接call B/C。

如果有了中介I,则流程如下:

# B and C ask I to be notified when A is done creating the resource

# ... # A is done creating the resource # A gets hold of the reference to the intermediary I # A calls I I->notify()

只要B/C订阅了router creation消息,A and ‘I’ 不需要知道B/C的状态和变化。

同时callback也可以抛出异常告诉A该操作是否可行,这就需要A先发一个events.BEFORE_CREATE来测试一下问问B/C的意见了。

实例代码如下:

from neutron.callbacks import events
from neutron.callbacks import exceptions
from neutron.callbacks import resources
from neutron.callbacks import registry


def callback1(resource, event, trigger, **kwargs):
    raise Exception('I am failing!')

def callback2(resource, event, trigger, **kwargs):
    print('Callback2 called by %s on event  %s' % (trigger, event))


registry.subscribe(callback1, resources.ROUTER, events.BEFORE_CREATE)
registry.subscribe(callback2, resources.ROUTER, events.BEFORE_CREATE)
registry.subscribe(callback2, resources.ROUTER, events.ABORT_CREATE)
print('Subscribed')


def do_notify():
    kwargs = {'foo': 'bar'}
    registry.notify(resources.ROUTER, events.BEFORE_CREATE, do_notify, **kwargs)


print('Notifying...')
try:
    do_notify()
except exceptions.CallbackFailure as e:
    print('Error: ', e)

Unsubscribing to events

  • clear(): it unsubscribes all subscribed callbacks: this can be useful especially when winding down the system, and notifications shall no longer be triggered.
  • unsubscribe(): it selectively unsubscribes a callback for a specific resource’s event. Say callback C has subscribed to event A for resource R, any notification of event A for resource R will no longer be handed over to C, after the unsubscribe() invocation.
  • unsubscribe_by_resource(): say that callback C has subscribed to event A, B, and C for resource R, any notification of events related to resource R will no longer be handed over to C, after the unsubscribe_by_resource() invocation.
  • unsubscribe_all(): say that callback C has subscribed to events A, B for resource R1, and events C, D for resource R2, any notification of events pertaining resources R1 and R2 will no longer be handed over to C, after the unsubscribe_all() invocation.

 git grep subscribe:

neutron/db/db_base_plugin_v2.py:            registry.subscribe(self.validate_network_rbac_policy_change,
neutron/db/external_net_db.py:        registry.subscribe(self._process_ext_policy_create,
neutron/db/external_net_db.py:            registry.subscribe(self._validate_ext_not_in_use_by_tenant,
neutron/db/l3_db.py:def subscribe():
neutron/db/l3_db.py:    registry.subscribe(
neutron/db/l3_db.py:    registry.subscribe(
neutron/db/l3_db.py:    registry.subscribe(
neutron/db/l3_db.py:    registry.subscribe(
neutron/db/l3_db.py:subscribe()
neutron/db/l3_dvrscheduler_db.py:def subscribe():
neutron/db/l3_dvrscheduler_db.py:    registry.subscribe(
neutron/db/l3_dvrscheduler_db.py:    registry.subscribe(
neutron/db/l3_dvrscheduler_db.py:    registry.subscribe(
neutron/objects/rbac_db.py:    def subscribe_to_rbac_events(class_instance):
neutron/objects/rbac_db.py:            registry.subscribe(class_instance.validate_rbac_policy_change,
neutron/objects/rbac_db.py:        mcs.subscribe_to_rbac_events(cls)
neutron/plugins/ml2/extensions/dns_integration.py:def subscribe():
neutron/plugins/ml2/extensions/dns_integration.py:    registry.subscribe(
neutron/plugins/ml2/extensions/dns_integration.py:    registry.subscribe(
neutron/plugins/ml2/extensions/dns_integration.py:    registry.subscribe(
neutron/plugins/ml2/extensions/dns_integration.py:subscribe()
neutron/services/auto_allocate/db.py:    registry.subscribe(_ensure_external_network_default_value_callback,
neutron/services/auto_allocate/db.py:    registry.subscribe(_ensure_external_network_default_value_callback,
neutron/services/auto_allocate/db.py:    registry.subscribe(_ensure_external_network_default_value_callback,
neutron/services/bgp/bgp_plugin.py:        registry.subscribe(self.floatingip_update_callback,
neutron/services/bgp/bgp_plugin.py:        registry.subscribe(self.router_interface_callback,
neutron/services/bgp/bgp_plugin.py:        registry.subscribe(self.router_interface_callback,
neutron/services/bgp/bgp_plugin.py:        registry.subscribe(self.router_interface_callback,
neutron/services/bgp/bgp_plugin.py:        registry.subscribe(self.router_gateway_callback,
neutron/services/bgp/bgp_plugin.py:        registry.subscribe(self.router_gateway_callback,
neutron/services/l3_router/l3_router_plugin.py:            l3_dvrscheduler_db.subscribe()
neutron/services/l3_router/l3_router_plugin.py:        l3_db.subscribe()

 比如neutron/db/l3_dvrscheduler_db.py: registry.subscribe(

在port delete之前要问问l3是否可以:

        _prevent_l3_port_delete_callback, resources.PORT, events.BEFORE_DELETE)

 l3的判断可以得条件是:

core plugin中没有它的记录

port的owner不属于router

没有fixed ip

owner router已经不存在

    def prevent_l3_port_deletion(self, context, port_id):
        """Checks to make sure a port is allowed to be deleted.

        Raises an exception if this is not the case.  This should be called by
        any plugin when the API requests the deletion of a port, since some
        ports for L3 are not intended to be deleted directly via a DELETE
        to /ports, but rather via other API calls that perform the proper
        deletion checks.
        """
        try:
            port = self._core_plugin.get_port(context, port_id)
        except n_exc.PortNotFound:
            # non-existent ports don't need to be protected from deletion
            return
        if port['device_owner'] not in self.router_device_owners:
            return
        # Raise port in use only if the port has IP addresses
        # Otherwise it's a stale port that can be removed
        fixed_ips = port['fixed_ips']
        if not fixed_ips:
            LOG.debug("Port %(port_id)s has owner %(port_owner)s, but "
                      "no IP address, so it can be deleted",
                      {'port_id': port['id'],
                       'port_owner': port['device_owner']})
            return
        # NOTE(kevinbenton): we also check to make sure that the
        # router still exists. It's possible for HA router interfaces
        # to remain after the router is deleted if they encounter an
        # error during deletion.
        # Elevated context in case router is owned by another tenant
        if port['device_owner'] == DEVICE_OWNER_FLOATINGIP:
            if not self._floating_ip_exists(context, port['device_id']):
                LOG.debug("Floating IP %(f_id)s corresponding to port "
                          "%(port_id)s no longer exists, allowing deletion.",
                          {'f_id': port['device_id'], 'port_id': port['id']})
                return
        elif not self._router_exists(context, port['device_id']):
            LOG.debug("Router %(router_id)s corresponding to port "
                      "%(port_id)s  no longer exists, allowing deletion.",
                      {'router_id': port['device_id'],
                       'port_id': port['id']})
            return

        reason = _('has device owner %s') % port['device_owner']
        raise n_exc.ServicePortInUse(port_id=port['id'],
                                     reason=reason)

 同时,在port delete之后,router需要发出 routers_updated notification:

    registry.subscribe(
        _notify_routers_callback, resources.PORT, events.AFTER_DELETE)
    def routers_updated(self, context, router_ids, operation=None, data=None,
                        shuffle_agents=False, schedule_routers=True):
        if router_ids:
            self._notification(context, 'routers_updated', router_ids,
                               operation, shuffle_agents, schedule_routers)

http://docs.openstack.org/developer/neutron/devref/callbacks.html

原文地址:https://www.cnblogs.com/allcloud/p/5484672.html