嗨，我正在使用docker + Flask + Gunicorn + SQLAlchemy构建RESTful服务。 但是，有时在执行SQL时会超时。
这是我的引擎：
def get_sybase_us_engine(self):
from spa_manager import spa_config
new_sybase_engine = create_engine(spa_config["SYBASE_URL"], connect_args=spa_config["SYBASE_CONN_ARGS"], echo=True, echo_pool=True, pool_size=10)
return new_sybase_engine
是的，sybase。
这是我使用引擎的方式：
import yaml
from flask import Blueprint, request
from flask_restful import Api, Resource
us_sql_test_blueprint = Blueprint('us_sql_test_blueprint', __name__)
us_sql_test_blueprint_api = Api(us_sql_test_blueprint)
class USSQLTest(Resource):
def post(self):
"""
SQL Test API
---
parameters:
- name: json_string
in: body
type: string
required: false
tags:
- SQL Test API
"""
from app.util.sql_engine import SQLEngine
from app.util.spa_logger import get_logger
logger = get_logger()
# get data from request
spa_data_str = request.get_data()
# convert to dict
requestbody = yaml.safe_load(spa_data_str)
logger.info("SQL TEST request incoming!")
logger.info("Req: %s" % str(requestbody))
# here I establish connection to DB
sql_engine = SQLEngine()
sybase_us_engine = sql_engine.get_sybase_us_engine()
sybase_us_conn = sybase_us_engine.connect()
logger.info("US connection established!")
from app.implement.spa_db_sql_test import us_sql_test
try:
# query data from DB
result_dict = us_sql_test(sybase_us_conn)
logger.info("No hang here, query successfully for all connections")
logger.info(str(result_dict))
return result_dict
except Exception, e:
logger.exception(e.message)
return e.message
finally:
# finnally close connection
logger.info("Close us connection")
sybase_us_conn.close()
sybase_us_engine.dispose()
us_sql_test_blueprint_api.add_resource(USSQLTest, '/sql-test-us')
但是在我的Gunicorn日志中，有一段时间工作超时（pid 265）。
[2018-07-23 17:24:24 +0000] [265] [DEBUG] POST /spa-automation/public/api/sql-test-us
[2018-07-23 17:24:24 +0000] [259] [DEBUG] POST /spa-automation/public/api/sql-test-ca
[2018-07-23 17:26:40 +0000] [301] [DEBUG] POST /spa-automation/public/api/sync-spa-info
[2018-07-23 17:26:40 +0000] [271] [DEBUG] POST /spa-automation/public/api/cust-refresh
[2018-07-23 17:29:24 +0000] [301] [DEBUG] POST /spa-automation/public/api/sql-test-local-mysql
[2018-07-23 17:29:24 +0000] [283] [DEBUG] POST /spa-automation/public/api/sql-test-mysql
[2018-07-23 17:29:24 +0000] [295] [DEBUG] POST /spa-automation/public/api/sql-test-ca
[2018-07-23 17:29:24 +0000] [289] [DEBUG] POST /spa-automation/public/api/sql-test-us
[2018-07-23 17:31:40 +0000] [289] [DEBUG] POST /spa-automation/public/api/cust-refresh
[2018-07-23 17:31:40 +0000] [283] [DEBUG] POST /spa-automation/public/api/sync-spa-info
[2018-07-23 17:34:24 +0000] [289] [DEBUG] POST /spa-automation/public/api/sql-test-mysql
[2018-07-23 17:34:24 +0000] [271] [DEBUG] POST /spa-automation/public/api/sql-test-us
[2018-07-23 17:34:24 +0000] [259] [DEBUG] POST /spa-automation/public/api/sql-test-ca
[2018-07-23 17:34:24 +0000] [295] [DEBUG] POST /spa-automation/public/api/sql-test-local-mysql
[2018-07-23 17:36:40 +0000] [289] [DEBUG] POST /spa-automation/public/api/sync-spa-info
[2018-07-23 17:36:40 +0000] [259] [DEBUG] POST /spa-automation/public/api/cust-refresh
[2018-07-23 17:39:24 +0000] [289] [DEBUG] POST /spa-automation/public/api/sql-test-us
[2018-07-23 17:39:24 +0000] [283] [DEBUG] POST /spa-automation/public/api/sql-test-ca
[2018-07-23 17:39:24 +0000] [295] [DEBUG] POST /spa-automation/public/api/sql-test-mysql
[2018-07-23 17:39:24 +0000] [277] [DEBUG] POST /spa-automation/public/api/sql-test-local-mysql
[2018-07-23 17:39:25 +0000] [7] [CRITICAL] WORKER TIMEOUT (pid:265)
这是我的项目日志：
[2018-07-23 17:24:25,425] - spa_test_us_view.py [Line:32] - [INFO]-[thread:140514956412672]-[process:265] - US connection established!
[2018-07-23 17:24:25,426] - log.py [Line:109] - [INFO]-[thread:140514956412672]-[process:265] -
select top 10 * from part_master
[2018-07-23 17:24:25,426] - log.py [Line:109] - [INFO]-[thread:140514956412672]-[process:265] - ()
此进程：265挂在那里，没有进程265的进一步日志
这是我在dockerfile中启动Gunicorn的方法：
ENTRYPOINT gunicorn -w 8 spa_manager:spa_app -b 0.0.0.0:80 --log-level=debug --timeout 900
顺便说一句，如果我不使用flask和Gunicorn，只需使用Python脚本，就不会发生超时。
出了什么问题？我不知道...