ORA-24550错误的原因是什么？

Question

1. python3.5.2，scrapy1.2.1，cx_Oracle5.2.1

2. 用scrapy抓取，然后存oracle，程序运行几分钟会崩溃

3. pipelines.py如下：

# -*- coding: utf-8 -*-

import os
import cx_Oracle
from twisted.enterprise import adbapi
from company.items import HaiguanItem
os.environ['NLS_LANG'] = 'SIMPLIFIED CHINESE_CHINA.UTF8'


class BasePiPeline(object):

    def __init__(self, dbargs, insert_sql):
        self.dbargs = dbargs
        self.insert_sql = insert_sql

    def open_spider(self, spider):
        self.dbpool = adbapi.ConnectionPool('cx_Oracle', **self.dbargs)

    def close_spider(self, spider):
        self.dbpool.close()

    def process_item(self, item, spider):
        self.dbpool.runInteraction(self._do_insert, item, spider)

    def _do_insert(self, txn, item, spider):
        try:
            txn.execute(self.insert_sql, dict(item))
            spider.log("保存数据成功： %s" % item['company_name'])
        except cx_Oracle.IntegrityError:
            spider.log("该条数据已存在： %s" % item['company_name'])
        except Exception as e:
            spider.log(str(e))

    @staticmethod
    def create_insert_sql(table, *columns):
        sql = "insert into {} ".format(table) + "(" + ", ".join(columns) + \
              ") values (:" + ", :".join(columns) + ")"
        return sql


class HaiguanPipeline(BasePiPeline):

    @classmethod
    def from_crawler(cls, crawler):
        dbargs = crawler.settings.get('DATABASES').get('oracle')
        table = 'company_custom_rating'
        columns = list(HaiguanItem.fields.keys())
        return cls(dbargs=dbargs,
                   insert_sql=cls.create_insert_sql(table, *columns))

Answer 1

问题已经解决，使用adbapi连接数据库时指定threaded=True，好像是关于线程安全的问题，具体原因不太清楚，附上ORA-24550文档解释ORA-24550