理解OpenStack+Ceph（5）：OpenStack与Ceph之间的集成[OpenStackIntegrationwithCeph]...

def get(self, location, offset&＃61;0, chunk_size&＃61;None, context&＃61;None):return (ImageIterator(loc.pool, loc.image, loc.snapshot, self), self.get_size(location))
1. 根据 location&＃xff0c;定位到 rbd image
2. 调用 image.read 方法&＃xff0c;按照 chunk size 读取 image data
3. 将 image data 返回调用端

def get_size(self, location, context&＃61;None):
1. 找到 store location&＃xff1a;loc &＃61; location.store_location
2. 使用 loc 中指定的 pool 或者配置的默认pool
3. 建立 connection 和 打开 IO Context&＃xff1a;with rbd.Image(ioctx, loc.image, snapshot&＃61;loc.snapshot) as image
4. 获取 image.stat() 并获取 “size”

def add(self, image_id, image_file, image_size, context&＃61;None):
1. 将 image_id 作为 rbd image name&＃xff1a;image_name &＃61; str(image_id)
2. 创建 rbd image&＃xff1a;librbd.create(ioctx, image_name, size, order, old_format&＃61;False, features&＃61;rbd.RBD_FEATURE_LAYERING)。如果 rbd 支持 RBD_FEATURE_LAYERING 的话&＃xff0c;创建一个 clonable snapshot&＃xff1a;rbd://fsid/pool/image/snapshot&＃xff1b;否则&＃xff0c;创建一个 rbd image&＃xff1a;rbd://image
3. 调用 image.write 方法将 image_file 的内容按照 chunksize 依次写入 rbd image
4. 如果要创建 snapshot 的话&＃xff0c;调用 image.create_snap(loc.snapshot) 和 image.protect_snap(loc.snapshot) 创建snapshot

def delete(self, location, context&＃61;None):
1. 根据 location 计算出 rbd image&＃xff0c;snapshot 和 pool
2. 如果它是一个 snapshot &＃xff08;location 是 rbd://fsid/pool/image/snapshot&＃xff09;&＃xff0c;则 unprotect_snap&＃xff0c;再 remove_snap&＃xff0c;然后删除 image&＃xff1b;这时候有可能出错&＃xff0c;比如存在基于该 image 的 volume。
3. 如果它不是一个 snapshot &＃xff08;location 是 rbd://image&＃xff09;&＃xff0c;直接删除 rbd image。这操作也可能会出错。

root&＃64;ceph1:~# rbd info images/71dc76da-774c-411f-a958-1b51816ec50f
rbd image &＃39;71dc76da-774c-411f-a958-1b51816ec50f&＃39;:
size 40162 kB in 5 objects
order 23 (8192 kB objects)
block_name_prefix: rbd_data.103d2246be0e
format: 2
features: layering

root&＃64;ceph1:~# rbd snap ls images/71dc76da-774c-411f-a958-1b51816ec50f
SNAPID NAME SIZE2 snap 40162 kB

root&＃64;ceph1:~# rbd info images/71dc76da-774c-411f-a958-1b51816ec50f&＃64;snap
rbd image &＃39;71dc76da-774c-411f-a958-1b51816ec50f&＃39;:
size 40162 kB in 5 objects
order 23 (8192 kB objects)
block_name_prefix: rbd_data.103d2246be0e
format: 2
features: layering
protected: True

root&＃64;ceph1:~# rbd children images/71dc76da-774c-411f-a958-1b51816ec50f&＃64;snap
volumes/volume-65dbaf38-0b9d-4654-bba4-53f12cc906e3
volumes/volume-6868a043-1412-4f6c-917f-bbffb1a8d21a

The image cannot be deleted because it is in use through the backend store outside of Glance.
这是因为该 image 的 rbd image 的 snapshot 被使用了。

def initialize_connection(self, volume, connector):
hosts, ports &＃61; self._get_mon_addrs() #调用 args &＃61; [&＃39;ceph&＃39;, &＃39;mon&＃39;, &＃39;dump&＃39;, &＃39;--format&＃61;json&＃39;] 获取 monmap&＃xff0c;再获取 hosts 和 ports
data &＃61; {&＃39;driver_volume_type&＃39;: &＃39;rbd&＃39;,&＃39;data&＃39;: {&＃39;name&＃39;: &＃39;%s/%s&＃39; % (self.configuration.rbd_pool, volume[&＃39;name&＃39;]), &＃39;hosts&＃39;: hosts, &＃39;ports&＃39;: ports, &＃39;auth_enabled&＃39;: (self.configuration.rbd_user is not None), &＃39;auth_username&＃39;: self.configuration.rbd_user, &＃39;secret_type&＃39;: &＃39;ceph&＃39;, &＃39;secret_uuid&＃39;: self.configuration.rbd_secret_uuid, } }

client &＃61; self.rados.Rados(rados_id&＃61;self.configuration.rbd_user, conffile&＃61;self.configuration.rbd_ceph_conf)
client.connect(timeout&＃61; self.configuration.rados_connect_timeout)
ioctx &＃61; client.open_ioctx(pool)

ioctx.close()
client.shutdown()