Chinaunix首页 | 论坛 | 博客

好好锻炼，一定要保重自己，身体重要啊---笨笨语录

首页　| 　博文目录　| 　关于我

satfire

  • 博客访问： 378817
  • 博文数量： 124
  • 博客积分： 30
  • 博客等级： 民兵
  • 技术积分： 11
  • 用 户 组： 普通用户
  • 注册时间： 2011-08-29 22:22
文章分类

全部博文（124）

文章存档

2016年（8）

2015年（52）

2014年（53）

2013年（11）

我的朋友
最近访客
推荐博文
相关博文
OpenStack 创建快照(Create Snapshot)过程源码分析

分类： 云计算

2015-04-17 15:31:10

原文地址：OpenStack 创建快照(Create Snapshot)过程源码分析 作者：liujunwei1234

1. 背景：
    一直以为OpenStack的创建快照的操作是在线创建快照（live snapshot）, 并且应该是增量的快照，即利用virsh或者qemu的live snapshot来实现的：
         virsh  snapshot-create-as --live ....
    后来发现快照和原始镜像之间并没有依赖关系，感觉OpenStack还做的挺好的，自动解决了增量快照和原始镜像之间的依赖关系;
    但是后来又发现做快照的时候虚拟机竟然会shutoff, 就感觉不对了，于是分析了下源码。

2. 结论：目前OpenStack默认的快照方式都是cold snapshot, 首先先关机，其次执行如下命令生成一个镜像文件，再次开机，最后再调用glance api将镜像上传。
       qemu-img convert -f qcow2 -O qcow2  
       所以目前并不是真正意义的快照，其实和关闭虚拟机，拷贝一份，再上传没有本质区别。

3. 源代码流程分析

3.1 nova/compute/api.py
  1. # NOTE(melwitt): We don't check instance lock for snapshot because lock is
  2.     # intended to prevent accidental change/delete of instances
  3.     @wrap_check_policy
  4.     @check_instance_cell
  5.     @check_instance_state(vm_state=[vm_states.ACTIVE, vm_states.STOPPED,
  6.                                     vm_states.PAUSED, vm_states.SUSPENDED])
  7.     def snapshot(self, context, instance, name, extra_properties=None):
  8.         """Snapshot the given instance.

  10.         :param instance: nova.db.sqlalchemy.models.Instance
  11.         :param name: name of the snapshot
  12.         :param extra_properties: dict of extra image properties to include
  13.                                  when creating the image.
  14.         :returns: A dict containing image metadata
  15.         """
  16.        #调用glance api创建image entry，为后将snapshot上传为镜像做准备，虽然镜像和snapshot在可以上传到glance作为镜像启动虚拟机，
  17.        #但是为了区分二者的不同，glance将镜像和snapshot标记卫不同的类型：type=image 和 type=snapshot
  18.         image_meta = self._create_image(context, instance, name,
  19.                                         'snapshot',
  20.                                         extra_properties=extra_properties)

  22.         # NOTE(comstud): Any changes to this method should also be made
  23.         # to the snapshot_instance() method in nova/cells/messaging.py
  24.         # 将任务状态(task state) 设置为:image_snapshot_pending
  25.         instance.task_state = task_states.IMAGE_SNAPSHOT_PENDING
  26.         instance.save(expected_task_state=[None])
  27.         #通过rpc调用nova/compute/rpcapi.py的snapshot_instance函数
  28.         self.compute_rpcapi.snapshot_instance(context, instance,
  29.                                               image_meta['id
3.2 nova/compute/rpcapi.py
  1. #梳理下流程： （1）用户发起create snapshot的请求; （2）nova-api服务接收到这个请求并进行前期处理，即3.1中代码的处理流程;
  2.                             （3）真正的snapshot操作是需要在nova-compute节点上执行的，所以nova-api需要向nova-compute发送message
  3. #由于OpenStack环境中会有很多个nova-compute,所以需要通过server=_compute_host(None, instance)来获取虚拟机所在的host，并向其发送message。
  4. def snapshot_instance(self, ctxt, instance, image_id):
  5.         version = '3.0'
  6.         cctxt = self.client.prepare(server=_compute_host(None, instance),
  7.                 version=version)
  8.         cctxt.cast(ctxt, 'snapshot_instance',
  9.                    instance=instance,
  10.                    image_id=image_id)
3.3 nova/virt/libvirt/driver.py
  1.  def snapshot(self, context, instance, image_id, update_task_state):
  2.         """Create snapshot from a running VM instance.

  4.         This command only works with qemu 0.14+
  5.         """
  6.         try:
  7.             virt_dom = self._get_domain(instance)
  8.         except exception.InstanceNotFound:
  9.             raise exception.InstanceNotRunning(instance_id=instance['uuid'])

  11.         base_image_ref = instance['image_ref']

  13.         base = compute_utils.get_image_metadata(
  14.             context, self._image_api, base_image_ref, instance)

  16.         snapshot = self._image_api.get(context, image_id)

  18.         disk_path = libvirt_utils.find_disk(virt_dom)
  19.         source_format = libvirt_utils.get_disk_type(disk_path)

  21.         image_format = CONF.libvirt.snapshot_image_format or source_format

  23.         # NOTE(bfilippov): save lvm and rbd as raw
  24.         if image_format == 'lvm' or image_format == 'rbd':
  25.             image_format = 'raw'

  27.         metadata = self._create_snapshot_metadata(base,
  28.                                                   instance,
  29.                                                   image_format,
  30.                                                   snapshot['name'])

  32.         snapshot_name = uuid.uuid4().hex

  34.         state = LIBVIRT_POWER_STATE[virt_dom.info()[0]]

  36.         # NOTE(rmk): Live snapshots require QEMU 1.3 and Libvirt 1.0.0.
  37.         # These restrictions can be relaxed as other configurations
  38.         # can be validated.
  39.         # NOTE(dgenin): Instances with LVM encrypted ephemeral storage require
  40.         # cold snapshots. Currently, checking for encryption is
  41.         # redundant because LVM supports only cold snapshots.
  42.         # It is necessary in case this situation changes in the
  43.         # future.
  44.         #这里需要注意，解释了为啥现在是cold snapshot而不是live snapshot:
  45.         # 有人提过live snapshot的bug，社区认为live snapshot目前不稳定，所以默认条件下采用cold snapshot，并且是通过硬编码来实现的
  46.         # 看下面这个判断条件，成立的时候将live_snapshot = true,其中MIN_LIBVIRT_LIVESNAPSHOT_VERSION=1.3.0, 其实现在libvirt的最新版本
  47.         # 才到1.2.11， 所以这个live_snapshot的条件不满足，就变成了cold_snapshot
  48.         if (self._host.has_min_version(MIN_LIBVIRT_LIVESNAPSHOT_VERSION,
  49.                                        MIN_QEMU_LIVESNAPSHOT_VERSION,
  50.                                        REQ_HYPERVISOR_LIVESNAPSHOT)
  51.              and source_format not in ('lvm', 'rbd')
  52.              and not CONF.ephemeral_storage_encryption.enabled):
  53.             live_snapshot = True
  54.             # Abort is an idempotent operation, so make sure any block
  55.             # jobs which may have failed are ended. This operation also
  56.             # confirms the running instance, as opposed to the system as a
  57.             # whole, has a new enough version of the hypervisor (bug 1193146).
  58.             try:
  59.                 virt_dom.blockJobAbort(disk_path, 0)
  60.             except libvirt.libvirtError as ex:
  61.                 error_code = ex.get_error_code()
  62.                 if error_code == libvirt.VIR_ERR_CONFIG_UNSUPPORTED:
  63.                     live_snapshot = False
  64.                 else:
  65.                     pass
  66.         else:
  67.             live_snapshot = False

  69.         # NOTE(rmk): We cannot perform live snapshots when a managedSave
  70.         # file is present, so we will use the cold/legacy method
  71.         # for instances which are shutdown.
  72.         if state == power_state.SHUTDOWN:
  73.             live_snapshot = False

  75.         # NOTE(dkang): managedSave does not work for LXC
  76.         #注意这里，如果live_snashot目前是false,所以在做snapshot之前先要执行：
  77.         #（1）_detach_pci_devices， 卸载虚拟机挂载的pci设备，比如数据盘
  78.         #（2）  self._detach_sriov_ports， 卸载虚拟机挂载的SRIOV设备，比如支持SRIOV的网卡设备
  79.         if CONF.libvirt.virt_type != 'lxc' and not live_snapshot:
  80.             if state == power_state.RUNNING or state == power_state.PAUSED:
  81.                 self._detach_pci_devices(virt_dom,
  82.                     pci_manager.get_instance_pci_devs(instance))
  83.                 self._detach_sriov_ports(instance, virt_dom)
  84.                 virt_dom.managedSave(0)
  85.         #判断虚拟机的后端存储是什么，不同的后端存储做snapshot是不同的，本地文件系统的化，默认qcow2
  86.         snapshot_backend = self.image_backend.snapshot(instance,
  87.                 disk_path,
  88.                 image_type=source_format)

  90.         if live_snapshot:
  91.             LOG.info(_LI("Beginning live snapshot process"),
  92.                      instance=instance)
  93.         else:
  94.             LOG.info(_LI("Beginning cold snapshot process"),
  95.                      instance=instance)
  96.         #更新任务的状态为：image_pending_upload, 大家都知道做完snapshot要上传
  97.         update_task_state(task_state=task_states.IMAGE_PENDING_UPLOAD)
  98.         #目前做快照的过程是：
  99.         #（1）现在../data/nova/instance/snapshots目录下生成临时目录，比如nova/instances/snapshots/tmptHr585
  100.         #（2）然后将快照生成到这个目录，具体参见snapshot_backend.snapshot_extract(out_path, image_format)这个函数
  101.         #（3）生成完成后，通过glance api上传，具体参见 self._image_api.update
  102.         snapshot_directory = CONF.libvirt.snapshots_directory
  103.         fileutils.ensure_tree(snapshot_directory)
  104.         with utils.tempdir(dir=snapshot_directory) as tmpdir:
  105.             try:
  106.                 out_path = os.path.join(tmpdir, snapshot_name)
  107.                 if live_snapshot:
  108.                     # NOTE(xqueralt): libvirt needs o+x in the temp directory
  109.                     os.chmod(tmpdir, 0o701)
  110.                     self._live_snapshot(context, instance, virt_dom, disk_path,
  111.                                         out_path, image_format, base)
  112.                 else:
  113.                     #这个函数实际执行了一条命令： qemu-img convert -f qcow2 -O qcow2  disk_path out_path,算是生成了快照
  114.                     snapshot_backend.snapshot_extract(out_path, image_format)
  115.             finally:
  116.                 new_dom = None
  117.                 # NOTE(dkang): because previous managedSave is not called
  118.                 # for LXC, _create_domain must not be called.
  119.                 if CONF.libvirt.virt_type != 'lxc' and not live_snapshot:
  120.                     if state == power_state.RUNNING:
  121.                         new_dom = self._create_domain(domain=virt_dom)   ##恢复做快照之前虚拟机的状态
  122.                     elif state == power_state.PAUSED:
  123.                         new_dom = self._create_domain(domain=virt_dom,
  124.                                 launch_flags=libvirt.VIR_DOMAIN_START_PAUSED)
  125.                     if new_dom is not None:
  126.                         self._attach_pci_devices(new_dom,
  127.                             pci_manager.get_instance_pci_devs(instance))
  128.                         self._attach_sriov_ports(context, instance, new_dom)
  129.                 LOG.info(_LI("Snapshot extracted, beginning image upload"),
  130.                          instance=instance)

  132.             # Upload that image to the image service

  134.             update_task_state(task_state=task_states.IMAGE_UPLOADING,
  135.                      expected_state=task_states.IMAGE_PENDING_UPLOAD)
  136.             with libvirt_utils.file_open(out_path) as image_file:       ###将生成的快照上传到glance
  137.                 self._image_api.update(context,
  138.                                        image_id,
  139.                                        metadata,
  140.                                        image_file)
  141.                 LOG.info(_LI("Snapshot image upload complete"),
  142.                          instance=instance)



------------------------------------------------------------
   email: ustc.dylan@gmail.com
   微博：@Marshal-Liu
------------------------------------------------------------


阅读(1431) | 评论(0) | 转发(0) |
0

上一篇：PostgreSQL数据库性能压力测试实验---pgbench

下一篇：openstack policy 鉴权过程分析

给主人留下些什么吧！~~
关于我们 | 关于IT168 | 联系方式 | 广告合作 | 法律声明 | 免费注册

Copyright 2001-2010 ChinaUnix.net All Rights Reserved 北京皓辰网域网络信息技术有限公司. 版权所有

感谢所有关心和支持过ChinaUnix的朋友们

16024965号-6