1.需求背景
系统程序突然报错，报错信息如下：
The transaction log for database '@dbname' is full. To find out why space in the log cannot be reused, see the log_reuse_wait_desc column in sys.databases
此时查看log文件，已达2T。
当时的紧急处理方案是，移除掉镜像，修改数据库恢复模式（由full修改为simple），收缩日志。
为了防止类似问题再次发生，需对log 文件的大小进行监控，当到达阈值后，触发告警。
2.主要基础组件（类）
配置文件 qqmssqltest_db_server_conf.ini
同过此配置文件获取DB Server信息、DB信息、UID信息、邮件服务器信息等。
[sqlserver] db_user = XXXXXX db_pwd = XXXXXXX [sqlserver_qq] db_host = 110.119.120.114 db_port = 1433 [windows] user = pwd = [mail] host = zheshiceshidemail.qq.com port = 25 user = pwd = sender = zhejiushiceshidebuyaodangzhen@qq.com
获取连接串的组件mssql_get_db_connect.py
# -*- coding: utf-8 -*- import sys import os import datetime import configparser import pymssql # pip3 install pymssql-2.1.4-cp37-cp37m-win_amd64.whl # pip3 install pymssql -i https://pypi.doubanio.com/simple # 获取连接串信息 def mssql_get_db_connect(db_host, db_port): db_host = db_host db_port = db_port db_ps_file = os.path.join(sys.path[0], "qqmssqltest_db_server_conf.ini") config = configparser.ConfigParser() config.read(db_ps_file, encoding="utf-8") db_user = config.get('sqlserver', 'db_user') db_pwd = config.get('sqlserver', 'db_pwd') conn = pymssql.connect(host=db_host, port=db_port, user=db_user, password=db_pwd, charset="utf8", login_timeout=5, timeout=600, autocommit=True) return conn
执行SQL语句的组件mysql_exec_sql.py
# -*- coding: utf-8 -*- import mysql_get_db_connect def mysql_exec_dml_sql(db_host, db_port, exec_sql): conn = mysql_get_db_connect.mysql_get_db_connect(db_host, db_port) with conn.cursor() as cursor_db: cursor_db.execute(exec_sql) conn.commit() def mysql_exec_select_sql(db_host, db_port, exec_sql): conn = mysql_get_db_connect.mysql_get_db_connect(db_host, db_port) with conn.cursor() as cursor_db: cursor_db.execute(exec_sql) sql_rst = cursor_db.fetchall() return sql_rst def mysql_exec_select_sql_include_colnames(db_host, db_port, exec_sql): conn = mysql_get_db_connect.mysql_get_db_connect(db_host, db_port) with conn.cursor() as cursor_db: cursor_db.execute(exec_sql) sql_rst = cursor_db.fetchall() col_names = cursor_db.description return sql_rst, col_names
发邮件的功能send_monitor_mail.py
# -*- coding: utf-8 -*- # pip3 install PyEmail import smtplib from email.mime.text import MIMEText import configparser import os import sys # 发送告警邮件 def send_monitor_mail(mail_subject, mail_body, mail_receivers="testwukongbaigujing@qq.com"): db_ps_file = os.path.join(sys.path[0], "qqmssqltest_db_server_conf.ini") config = configparser.ConfigParser() config.read(db_ps_file, encoding="utf-8") mail_host = config.get('mail', 'host') mail_port = config.get('mail', 'port') # mail_user = config.get('mail', 'user') # mail_pwd = config.get('mail', 'pwd') sender = config.get('mail', 'sender') # receivers = config.get('mail', 'receivers') # 发送HTML格式邮件 message = MIMEText(mail_body, 'html', 'utf-8') # message = MIMEText(mail_body, 'plain', 'utf-8') message['subject'] = mail_subject message['From'] = sender message['To'] = mail_receivers try: smtpObj = smtplib.SMTP() smtpObj.connect(mail_host, mail_port) # 25 为 SMTP 端口号 # SMTP AUTH extension not supported by server. # https://github.com/miguelgrinberg/microblog/issues/76 # smtpObj.ehlo() # smtpObj.starttls() # smtpObj.login(mail_user, mail_pwd) smtpObj.sendmail(sender, mail_receivers, message.as_string()) smtpObj.quit() print("邮件发送成功") except Exception as e: print(e) # except smtplib.SMTPException: # print("Error: 无法发送邮件")
3.主要功能代码
收集到的DB数据文件的信息保存到表mssql_dblogsize中，其建表的脚本如下：
CREATE TABLE [dbo].[mssql_dblogsize]( [id] [int] IDENTITY(1,1) NOT NULL, [createtime] [datetime] NULL, [vip] [nvarchar](100) NULL, [port] [nvarchar](100) NULL, [Environment] [nvarchar](200) NULL, [Dbname] [varchar](200) NULL, [Logical_Name] [varchar](200) NULL, [Physical_Name] [varchar](1500) NULL, [Size] [bigint] NULL, PRIMARY KEY CLUSTERED ( [id] ASC )WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY] ) ON [PRIMARY] GO ALTER TABLE [dbo].[mssql_dblogsize] ADD DEFAULT (getdate()) FOR [createtime] GO
为了方便对表mssql_dblogsize的数据进行管理和展示，在其基础上抽象加工出了一个视图v_mssql_dblogsize，注意Size大小的转换（Size/128/1024 as SizeGB）
创建视图的脚本如下：
CREATE view [dbo].[v_mssql_dblogsize] as SELECT [id] ,[createtime] ,[vip] ,[port] ,[Environment] ,[Dbname] ,[Logical_Name] ,[Physical_Name] ,Size/128/1024 as SizeGB FROM [dbo].[mssql_dblogsize] where size >50*128*1024 and Physical_Name like '%ldf%' GO
本测试实例使用的数据库为qqDB，监控的各个DB Server保存在了表QQDBServer中，注意Port 不一定为标准端口1433.
collect_mssql_dblogsize_info.py
# -*- coding: utf-8 -*- import sys import os import configparser import pymssql import mssql_get_db_connect import mssql_exec_sql from datetime import datetime def collect_mssql_dblogsize_info(): db_ps_file = os.path.join(sys.path[0], "qqmssqltest_db_server_conf.ini") config = configparser.ConfigParser() config.read(db_ps_file, encoding="utf-8") m_db_host = config.get('sqlserver_qq', 'db_host') m_db_port = config.getint('sqlserver_qq', 'db_port') # 获取需要遍历的DB列表 exec_sql_1 = """ SELECT IP, case Port when '1444,1433' then '1433' else Port end as Port, Environment FROM qqDB.dbo.QQDBServer where InUse =1 AND ServerType IN ('SQL') and IP=VIP ; """ sql_rst_1 = mssql_exec_sql.mssql_exec_select_sql(m_db_host, m_db_port, exec_sql_1) for j in sql_rst_1: db_host_2 = j[0] db_port_2 = j[1] db_Environment = j[2] exec_sql_2 = """ select '""" + db_host_2 + """' as vip, '""" + db_port_2 + """' as port, '""" + db_Environment + """' as Environment,DB_NAME(database_id) AS DatabaseName, Name AS Logical_Name, Physical_Name, size FROM master.sys.master_files; """ try: sql_rst_2 = mssql_exec_sql.mssql_exec_select_sql(db_host_2, db_port_2, exec_sql_2) except Exception as e: print(e) for k in sql_rst_2: exec_sql_3 = """ insert into qqDB..mssql_dblogsize([vip], [port], [Environment], [Dbname], [Logical_Name], [Physical_Name], [Size]) values('%s', '%s', '%s', '%s', '%s', '%s', '%s'); """ conn = mssql_get_db_connect.mssql_get_db_connect(m_db_host, m_db_port) with conn.cursor() as cursor_db: cursor_db.execute(exec_sql_3 % (k[0], k[1], k[2], k[3], k[4], k[5], k[6] )) conn.commit() collect_mssql_dblogsize_info()
告警邮件的功能实现为mssql_alert_dblogsize.py,此份代码的告警阈值设置的为50G，数据来自于视图v_mssql_dblogsize。
# -*- coding: utf-8 -*- import sys import os import configparser import pymssql import mssql_get_db_connect import mssql_exec_sql import datetime import send_monitor_mail import pandas as pd def mssql_alert_dblogsize(): mail_subject = "SQL Server DB Log Size Greater than 50G, please check!!! " mail_receivers = "testDBAgrp@qtiantianq.com" db_ps_file = os.path.join(sys.path[0], "qqmssqltest_db_server_conf.ini") config = configparser.ConfigParser() config.read(db_ps_file, encoding="utf-8") m_db_host = config.get('sqlserver_qq', 'db_host') m_db_port = config.getint('sqlserver_qq', 'db_port') # 获取需要遍历的DB列表 exec_sql_4 = """ SELECT [vip] as IP,[port],[Environment],[Dbname] ,[Logical_Name],[Physical_Name],[SizeGB],[createtime] FROM qqDB.[dbo].[v_mssql_dblogsize] order by VIP,Dbname; """ sql_rst_4, col_name = mssql_exec_sql.mssql_exec_select_sql_include_colnames(m_db_host, m_db_port, exec_sql_4) # print(sql_rst_4) if len(sql_rst_4): mail_time = datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S') columns = [] for i in range(len(col_name)): columns.append(col_name[i][0]) df = pd.DataFrame(columns=columns) for i in range(len(sql_rst_4)): df.loc[i] = list(sql_rst_4[i]) mail_body = df.to_html(index=False, justify="left").replace('<th>', '<th style = "color:red; text-align:left; background-color: yellow">') mail_html = "<html><body><h4>" + "Deal All : " + "<br><h4>" + "以下数据库的db log文件，已大于50G.请及时检查，谢谢！ " + "<br><h4>" + mail_body + "</body></html>" send_monitor_mail.send_monitor_mail(mail_subject=mail_subject, mail_body=mail_html, mail_receivers=mail_receivers) mssql_alert_dblogsize()
4.实现
定时任务是通过windows的计划任务来实现的,在此不做过多的叙述。告警邮件的部分截图如下:
5.附录
1.报错定位，判断是不是log文件过大
https://blog.csdn.net/weixin_30785593/article/details/99912405
2.关于为什么数据库log文件过大，我们可以参考以下分享的文章