简介:
相对于最新的MySQL5.6,MariaDB在性能、功能、管理、NoSQL扩展方面包含了更丰富的特性。比如微秒的支持、线程池、子查询优化、组提交、进度报告等。
本文就主要探索MariaDB当中连接池的一些特性,配置。来配合我们的sqlalchemy。
一:起因
本来是不会写这个东西的,但是,写好了python--flask程序,使用sqlalchemy+mariadb,部署以后总是出问题,500错误之类的。
使用默认连接参数
engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan',)
错误提示是:
sqlalchemy.exc.OperationalError: (mysql.connector.errors.OperationalError) MySQL Connection not available. [SQL: 'SELECT public.id AS public_id, public.public_name AS public_public_name, public.public_email AS public_public_email \nFROM public \nWHERE public.public_name = %(public_name_1)s \n LIMIT %(param_1)s'] [parameters: [{}]] (Background on this error at: http://sqlalche.me/e/e3q8)
http://sqlalche.me/e/e3q8:
OperationalError:
Exception raised for errors that are related to the database’s operation andnot necessarily under the control of the programmer, e.g. an unexpecteddisconnect occurs, the data source name is not found, a transaction could notbe processed, a memory allocation error occurred during processing, etc.
This error is aDBAPI Errorand originates fromthe database driver (DBAPI), not SQLAlchemy itself.
TheOperationalErroris the most common (but not the only) error class usedby drivers in the context of the database connection being dropped, or notbeing able to connect to the database. For tips on how to deal with this, seethe sectionDealing with Disconnects.
意思是没有正确断开和数据库的连接。
二:处理断开
http://docs.sqlalchemy.org/en/latest/core/pooling.html#pool-disconnects
官方给了三种方案来解决这个问题:
1.悲观处理
engine = create_engine("mysql+pymysql://user:pw@host/db", pool_pre_ping=True)
pool_pre_ping=True
表示每次连接从池中检查,如果有错误,监测为断开的状态,连接将被立即回收。
2.自定义悲观的ping
from sqlalchemy importexcfrom sqlalchemy importeventfrom sqlalchemy importselect
some_engine=create_engine(...)
@event.listens_for(some_engine,"engine_connect")defping_connection(connection, branch):ifbranch:#"branch" refers to a sub-connection of a connection,
#we don't want to bother pinging on these.
return
#turn off "close with result". This flag is only used with
#"connectionless" execution, otherwise will be False in any case
save_should_close_with_result =connection.should_close_with_result
connection.should_close_with_result=Falsetry:#run a SELECT 1. use a core select() so that
#the SELECT of a scalar value without a table is
#appropriately formatted for the backend
connection.scalar(select([1]))exceptexc.DBAPIError as err:#catch SQLAlchemy's DBAPIError, which is a wrapper
#for the DBAPI's exception. It includes a .connection_invalidated
#attribute which specifies if this connection is a "disconnect"
#condition, which is based on inspection of the original exception
#by the dialect in use.
iferr.connection_invalidated:#run the same SELECT again - the connection will re-validate
#itself and establish a new connection. The disconnect detection
#here also causes the whole connection pool to be invalidated
#so that all stale connections are discarded.
connection.scalar(select([1]))else:raise
finally:#restore "close with result"
connection.should_close_with_result = save_should_close_with_result
说实话,没怎么看明白。
像是try一个select 语句,如果没问题就关闭。
3.乐观处理
from sqlalchemy importcreate_engine, exc
e=create_engine(...)
c=e.connect()try:#suppose the database has been restarted.
c.execute("SELECT * FROM table")
c.close()exceptexc.DBAPIError, e:#an exception is raised, Connection is invalidated.
ife.connection_invalidated:print("Connection was invalidated!")#after the invalidate event, a new connection#starts with a new Pool
c =e.connect()
c.execute("SELECT * FROM table")
这个看懂了,try一个select语句,如果无效,就返回Connection was invalidated!,然后开一个新的连接,再去执行select。这个应该写个装饰器,放在每个查询前面。
4.使用连接池回收
from sqlalchemy importcreate_engine
e= create_engine("mysql://scott:tiger@localhost/test", pool_recycle=3600)
这种方式就比较简单了,在连接参数中写上连接超时时间即可。
5.这是自己看文档找到的方法
from sqlalchemy.pool import QueuePool,NullPool,AssertionPool,StaticPool,SingletonThreadPool,Pool
在sqlalchemy.pool下有已经配置好的连接池,直接使用这些连接池也应该可以。
三:测试
docker run --restart=always --privileged --name My_mariadb_01 -p 3301:3306 -e MYSQL_ROOT_PASSWORD=123456 -d mariadb:10.2.13
docker run --restart=always --privileged --name My_mariadb_02 -p 3302:3306 -e MYSQL_ROOT_PASSWORD=123456 -d mariadb:10.2.13
docker run --restart=always --privileged --name My_mariadb_03 -p 3303:3306 -e MYSQL_ROOT_PASSWORD=123456 -d mariadb:10.2.13
docker run --restart=always --privileged --name My_mariadb_04 -p 3304:3306 -e MYSQL_ROOT_PASSWORD=123456 -d mariadb:10.2.13
docker run --restart=always --privileged --name My_mariadb_05 -p 3305:3306 -e MYSQL_ROOT_PASSWORD=123456 -d mariadb:10.2.13
为避免因数据库交叉连接,首先开启5个MARIADB
Flask_Plan_01 8801 engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan',)
Flask_Plan_02 8802 engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan', pool_pre_ping=True)
Flask_Plan_03 8803 engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan', poolclass=QueuePool)
Flask_Plan_04 8804 engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan', poolclass=NullPool)
Flask_Plan_05 8805 engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan', pool_recycle=3600)
用这5种连接参数进行连接测试。
如果你愿意,也可以继续开,QueuePool,NullPool,AssertionPool,StaticPool,SingletonThreadPool,Pool,把这几种都测试一下。
8801 8805 均会不同程度的出现500错误,8801频率还高点。
sqlalchemy.exc.OperationalError: (mysql.connector.errors.OperationalError) MySQL Connection not available. [SQL: 'SELECT public.id AS public_id, public.public_name AS public_public_name, public.public_email AS public_public_email \nFROM public \nWHERE public.public_name = %(public_name_1)s \n LIMIT %(param_1)s'] [parameters: [{}]] (Background on this error at: http://sqlalche.me/e/e3q8)
sqlalchemy.exc.OperationalError: (mysql.connector.errors.OperationalError) MySQL Connection not available. [SQL: 'SELECT public.id AS public_id, public.public_name AS public_public_name, public.public_email AS public_public_email \nFROM public \nWHERE public.public_name = %(public_name_1)s \n LIMIT %(param_1)s'] [parameters: [{}]] (Background on this error at: http://sqlalche.me/e/e3q8)
Internal Server Error
The server encountered an internal error and was unable to complete your request. Either the server is overloaded or there is an error in the application.
等会儿看看8802 8803 8804如何。
四:深入研究sqlalchemy源码
VENV\Flask_Base\Lib\site-packages\sqlalchemy\engine\__init__.py
看起来,没有默认值。所以engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan',)报错频率比较高。
五:研究pool源码
VENV\Flask_Base\Lib\site-packages\sqlalchemy\pool.py
看来poolclass的类型都定义在这里了。
1.SingletonThreadPool
A Pool that maintains one connection per thread
每个线程维护一个连接的池。
2.QueuePool
A :class:`.Pool` that imposes a limit on the number of open connections.
这种方式限制了连接数量,QueuePool是默认的连接池方式,除非使用了方言,也就是第三方链接库。
难怪我使用MySQL-connector-python时老出错呢,没打开连接池啊。
3.NullPool
A Pool which does not pool connections...
不使用连接池
4.StaticPool
A Pool of exactly one connection, used for all requests.
一个完整的连接池,用于所有的连接。
5.AssertionPool
A :class:`.Pool` that allows at most one checked out connection at any given time.
任何时间只给一个签出连接?为了debug模式?不懂了。
看的官方说明也没这么详细。
这么看来,如果我使用默认链接库,可以不加参数试试。
mysql-python是sqlalchemy默认的mysql链接库,我在windows下装不上。放弃测试默认链接库,手动指定连接池为QueuePool。
或者指定连接池类型为:QueuePool StaticPool SingletonThreadPool(多线程的时候)
六:连接池类型测试
修改测试docker
docker run --restart=always --privileged --name My_mariadb_01 -p 3301:3306 -e MYSQL_ROOT_PASSWORD=123456 -d mariadb:10.2.13
docker run --restart=always --privileged --name My_mariadb_02 -p 3302:3306 -e MYSQL_ROOT_PASSWORD=123456 -d mariadb:10.2.13
docker run --restart=always --privileged --name My_mariadb_03 -p 3303:3306 -e MYSQL_ROOT_PASSWORD=123456 -d mariadb:10.2.13
docker run --restart=always --privileged --name My_mariadb_04 -p 3304:3306 -e MYSQL_ROOT_PASSWORD=123456 -d mariadb:10.2.13
docker run --restart=always --privileged --name My_mariadb_05 -p 3305:3306 -e MYSQL_ROOT_PASSWORD=123456 -d mariadb:10.2.13
docker run --restart=always --privileged --name My_mariadb_06 -p 3306:3306 -e MYSQL_ROOT_PASSWORD=123456 -d mariadb:10.2.13
Flask_Plan_01 8801 engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan', pool_pre_ping=True))
Flask_Plan_02 8802 engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan', poolclass=SingletonThreadPool)
Flask_Plan_03 8803 engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan', poolclass=QueuePool)
Flask_Plan_04 8804 engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan', poolclass=NullPool)
Flask_Plan_05 8805 engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan', poolclass=StaticPool)
Flask_Plan_06 8806 engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan', poolclass=AssertionPool)
七:编写测试脚本
importrequestsimporttimei= 1
whileTrue:try:
r=requests.get('http://192.168.0.104:8801',timeout=5)if r.status_code==200:print(time.strftime('%Y-%m-%d %H:%M:%S')+'---'+str(i)+'---'+str(r.status_code)+'---ok')else:print(time.strftime('%Y-%m-%d %H:%M:%S') + '---' + str(i) + '---' + str(r.status_code) + '-----------badr')breaktime.sleep(1)
i+=1
except:print('except')print(time.strftime('%Y-%m-%d %H:%M:%S') +'---'+str(i)+'-----------bad')break
修改地址,把几个测试服务都开始跑。
出错就会停了。
代码很烂,凑活测试而已。
从晚上22:30睡觉到早上6:10起床,pool_pre_ping=True,SingletonThreadPool,QueuePool,NullPool,StaticPool,AssertionPool,都很稳定,访问代码都是200
八:继续研究相关代码
使用连接池进行多重处理
api文档--连接池的实现
classsqlalchemy.pool.Pool(creator,recycle=-1,echo=None,use_threadlocal=False,logging_name=None,reset_on_return=True,listeners=None,events=None,dialect=None,pre_ping=False,_dispatch=None)
Parameters:
creator–可调用的函数返回对象。
recycle– 超时回收时间。如果连接超过这个时间,连接就被关闭,换一个新的连接
logging_name - 日志标识名称
echo– 是否打印sql语句
use_threadlocal–是否使用线程,在同一应用程序的线程使用相同的连接对象
reset_on_return–在返回前的操作
rollback,大概是自动回滚
True 同为回滚
commit 大概是自动提交的意思
None 无操作
none 无操作
False 无操作
events– 列表元组,每个表单会传递给listen………………没搞懂
listeners - 弃用,被listen取代
dialect–链接库,使用create_engine时不使用,由引擎创建时处理
pre_ping–是否测试连接
基本上这些参数都在engine-creation-api中
Pool (creator,recycle=-1,echo=None,use_threadlocal=False,logging_name=None,reset_on_return=True,listeners=None,events=None,dialect=None,pre_ping=False,_dispatch=None)
StaticPool (creator,recycle=-1,echo=None,use_threadlocal=False,logging_name=None,reset_on_return=True,listeners=None,events=None,dialect=None,pre_ping=False,_dispatch=None)
NullPool (creator,recycle=-1,echo=None,use_threadlocal=False,logging_name=None,reset_on_return=True,listeners=None,events=None,dialect=None,pre_ping=False,_dispatch=None)
QueuePool (creator,pool_size=5,max_overflow=10,timeout=30,**kw)
SingletonThreadPool(creator,pool_size=5,**kw)
AssertionPool (*args,**kw)
这下清楚了,Pool,StaicPool,NullPool,都一样,直接回收,效率一定低了。
我们就指定默认的QueuePool好了。以后观察着服务器的负载,负载大了以后,调整就好了。
自定义方法如下:
engine = create_engine('mysql+mysqlconnector://plan:plan@mysql/plan',pool_size=5,
max_overflow=10,
pool_timeout=30,
pool_pre_ping=True)
九:总结
曲折的道路,终于找到了解决方案。
sqlalchemy的教程当中,很少有讲如何部署的。很多又是linux开发。可能在linux下很容易装默认链接库,部署的时候就自动使用了QueuePool连接池。所以这种问题很少出现。
我在windows下开发,部署在linux,开发和部署都使用了非默认链接库,导致没有使用默认连接池。
那么随着深入研究,找到了连接池的配置,并掌握这一知识,为以后的开发部署工作,扫除了障碍。
虽然源码里面还有很多看不懂,但是读书百遍其义自见,还是要多读(我是懒蛋,遇到问题,再去解决,下一个问题是什么呢?)。