nova quota

quota中文叫配额，是云计算中重要的概念之一，云计算提供什么，按需提供服务(各种资源)，按需是云的精髓，要多少给多少，超限的响应也不能给。

在openstack中nova随处可见，trace openstack的源码，安全组，网络数量，tenant的cpu，内存，磁盘，等等都有quota的身影。

quota的变动是一个数据库事务，包含reservation/commit/rollback三个动作，缺一不可，reserve时若发现超限，则commit不进行，正常则commit，失败

就rollback，数据库写事务时的常规套路。需要特别提一句的是，reserve成功时候会返回reservation_id的list，commit根据这些id进行对应的配额变化。

trace过很多次openstack创建instance 的过程，再一次：

在_provision_instance中_check_num_instances_quota：

quotas = objects.Quotas(context)
quotas.reserve(instances=max_count, cores=req_cores, ram=req_ram)

调用nova/objects/quotas.py，这里objects中的quota封装了reserve和commit过程中serveration_id的list，省得传递它了。

在objects/quotas.py中的Quota 类中可以看到：

 @base.remotable
    def reserve(self, expire=None, project_id=None, user_id=None,
                **deltas):
        reservations = quota.QUOTAS.reserve(self._context, expire=expire,
                                            project_id=project_id,
                                            user_id=user_id,
                                            **deltas)
        self.reservations = reservations
        self.project_id = project_id
        self.user_id = user_id
        self.obj_reset_changes()

这里quota.QUOTAS是QUOTAS = QuotaEngine()

一般来说，quota engine中对应的engine driver是DB，不用DB，难道用什么别的更合适？这里quota还要搞driver，想太多？

既然quota engine对应的是数据库，那就回到前面提到的sqlalchemy中去了，在nova/db/sqlalchemy/api.py中，我们可以找到

quota_reserve， reservation_commit，reservation_rollback等方法，背后对应了数据库的操作。

进入openstack的数据库中，查看表quotas, quota_usages, reservations几张表(还有别的表)就可以大致了解了。

比如quota表：

MariaDB [nova]> select * from quotas;
+----+---------------------+------------+------------+----------------------------------+-----------------------------+------------+---------+
| id | created_at          | updated_at | deleted_at | project_id                       | resource                    | hard_limit | deleted |
+----+---------------------+------------+------------+----------------------------------+-----------------------------+------------+---------+
|  1 | 2015-04-28 07:55:58 | NULL       | NULL       | 91accf025304465fa1a4cfdcaa94d91d | metadata_items              |        128 |       0 |
|  2 | 2015-04-28 07:55:58 | NULL       | NULL       | 91accf025304465fa1a4cfdcaa94d91d | injected_file_content_bytes |      10240 |       0 |
|  3 | 2015-04-28 07:55:58 | NULL       | NULL       | 91accf025304465fa1a4cfdcaa94d91d | ram                         |      51200 |       0 |
|  4 | 2015-04-28 07:55:58 | NULL       | NULL       | 91accf025304465fa1a4cfdcaa94d91d | instances                   |         10 |       0 |
|  5 | 2015-04-28 07:55:58 | NULL       | NULL       | 91accf025304465fa1a4cfdcaa94d91d | injected_files              |          5 |       0 |
|  6 | 2015-04-28 07:55:58 | NULL       | NULL       | 91accf025304465fa1a4cfdcaa94d91d | cores                       |         20 |       0 |
|  7 | 2015-06-01 12:16:44 | NULL       | NULL       | 51b4379ea8734aaca96cb724a2d7200f | metadata_items              |        128 |       0 |
|  8 | 2015-06-01 12:16:44 | NULL       | NULL       | 51b4379ea8734aaca96cb724a2d7200f | injected_file_content_bytes |      10240 |       0 |
|  9 | 2015-06-01 12:16:44 | NULL       | NULL       | 51b4379ea8734aaca96cb724a2d7200f | ram                         |      51200 |       0 |
| 10 | 2015-06-01 12:16:44 | NULL       | NULL       | 51b4379ea8734aaca96cb724a2d7200f | instances                   |         10 |       0 |
| 11 | 2015-06-01 12:16:44 | NULL       | NULL       | 51b4379ea8734aaca96cb724a2d7200f | injected_files              |          5 |       0 |
| 12 | 2015-06-01 12:16:44 | NULL       | NULL       | 51b4379ea8734aaca96cb724a2d7200f | cores                       |         30 |       0 |
+----+---------------------+------------+------------+----------------------------------+-----------------------------+------------+---------+

quota_usages 表:

MariaDB [nova]> select * from quota_usages;
+---------------------+---------------------+------------+----+----------------------------------+-----------------+--------+----------+---------------+---------+----------------------------------+
| created_at          | updated_at          | deleted_at | id | project_id                       | resource        | in_use | reserved | until_refresh | deleted | user_id                          |
+---------------------+---------------------+------------+----+----------------------------------+-----------------+--------+----------+---------------+---------+----------------------------------+
| 2015-04-28 08:05:18 | 2015-04-28 09:11:05 | NULL       |  1 | 51b4379ea8734aaca96cb724a2d7200f | instances       |      1 |        0 |          NULL |       0 | 60ce7a8cab9c477895ef37440b1d9c57 |
| 2015-04-28 08:05:18 | 2015-04-28 09:11:05 | NULL       |  2 | 51b4379ea8734aaca96cb724a2d7200f | ram             |    512 |        0 |          NULL |       0 | 60ce7a8cab9c477895ef37440b1d9c57 |
| 2015-04-28 08:05:18 | 2015-04-28 09:11:05 | NULL       |  3 | 51b4379ea8734aaca96cb724a2d7200f | cores           |      1 |        0 |          NULL |       0 | 60ce7a8cab9c477895ef37440b1d9c57 |
| 2015-04-28 08:05:18 | 2015-04-28 08:05:18 | NULL       |  4 | 51b4379ea8734aaca96cb724a2d7200f | security_groups |      1 |        0 |          NULL |       0 | 60ce7a8cab9c477895ef37440b1d9c57 |

at last： 如果想在nova之外的地方使用quota，当然是novaclient，看nova client 源码，在novaclient/v1_1/shell.py中，有方法do_quota_update，参数有哪些

也可以看到，因此调用rest call或者nova client都可以达到目的。