nova源码分析--创建虚拟机（5）

最新推荐文章于 2022-06-23 11:00:38 发布

lemonwsx

最新推荐文章于 2022-06-23 11:00:38 发布

阅读量938

点赞数 1

CC 4.0 BY-SA版权

分类专栏： openstack

本文链接：https://blog.youkuaiyun.com/lemonwsx/article/details/84789262

本文详细解析了Nova在创建虚拟机时的调度过程，从`schedule_and_build_instances`开始，逐步剖析了如何选择物理节点，涉及scheduler_utils、ComputeTaskManager、FilterScheduler等关键组件。调度过程包括过滤器的选择、权重计算以及资源声明，展示了如何根据配置文件调整过滤器和度量标准来优化调度策略。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

self.compute_task_api.schedule_and_build_instances函数的定义在nova/conductor/api.py中的类ComputeTaskAPI中，代码如下：

    def schedule_and_build_instances(self, context, build_requests,
                                     request_spec, image,
                                     admin_password, injected_files,
                                     requested_networks, block_device_mapping,
                                     tags=None):
        self.conductor_compute_rpcapi.schedule_and_build_instances(
            context, build_requests, request_spec, image,
            admin_password, injected_files, requested_networks,
            block_device_mapping, tags)

直接调用self.conductor_compute_rpcapi.schedule_and_build_instances函数，该函数在文件nova/conductor/rpcapi.py中的ComputeTaskAPI类中定义，代码如下：

    def schedule_and_build_instances(self, context, build_requests,
                                     request_specs,
                                     image, admin_password, injected_files,
                                     requested_networks,
                                     block_device_mapping,
                                     tags=None):
        version = '1.17'
        kw = {'build_requests': build_requests,
              'request_specs': request_specs,
              'image': jsonutils.to_primitive(image),
              'admin_password': admin_password,
              'injected_files': injected_files,
              'requested_networks': requested_networks,
              'block_device_mapping': block_device_mapping,
              'tags': tags}

        if not self.client.can_send_version(version):
            version = '1.16'
            del kw['tags']

        cctxt = self.client.prepare(version=version)
        cctxt.cast(context, 'schedule_and_build_instances', **kw)

准备参数kw，通过rpc调用服务端的schedule_and_build_instances函数。服务端对应的函数在文件nova/conductor/manager.py中的类ComputeTaskManager中，代码如下：

    def schedule_and_build_instances(self, context, build_requests,
                                     request_specs, image,
                                     admin_password, injected_files,
                                     requested_networks, block_device_mapping,
                                     tags=None):
        # Add all the UUIDs for the instances
        instance_uuids = [spec.instance_uuid for spec in request_specs]
        try:
            hosts = self._schedule_instances(context, request_specs[0],
                                             instance_uuids)
        except Exception as exc:
            LOG.exception('Failed to schedule instances')
            self._bury_in_cell0(context, request_specs[0], exc,
                                build_requests=build_requests,
                                block_device_mapping=block_device_mapping)
            return

self._schedule_instances函数的功能为为将要创建的虚拟机挑选运行的物理节点，代码定义如下：

    def _schedule_instances(self, context, request_spec,
                            instance_uuids=None):
        scheduler_utils.setup_instance_group(context, request_spec)
        hosts = self.scheduler_client.select_destinations(context,
            request_spec, instance_uuids)
        return hosts

scheduler_utils.setup_instance_group函数的定义在文件./nova/scheduler/utils.py中，获取group相关的信息，并设置到request_spec.instance_group中的变量。self.scheduler_client.select_destinations通过一系列的包装调用，最后将调用./nova/scheduler/rpcapi.py文件中的类SchedulerAPI中的select_destinations函数，定义如下：