通過Python實現對SQL Server 數據文件大小的監控告警功能

Posted on by

1.需求背景

系統程序突然報錯，報錯信息如下：

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

此時查看log文件，已達2T。

當時的緊急處理方案是，移除掉鏡像，修改數據庫恢復模式（由full修改為simple），收縮日志。

為瞭防止類似問題再次發生，需對log 文件的大小進行監控，當到達閾值後，觸發告警。

2.主要基礎組件（類）

配置文件qqmssqltest_db_server_conf.ini

同過此配置文件獲取DB Server信息、DB信息、UID信息、郵件服務器信息等。

[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 = [email protected]

獲取連接串的組件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


# 獲取連接串信息
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

執行SQL語句的組件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

發郵件的功能send_monitor_mail.py

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


# pip3 install PyEmail
import smtplib
from email.mime.text import MIMEText
import configparser
import os
import sys


# 發送告警郵件
def send_monitor_mail(mail_subject, mail_body, mail_receivers="[email protected]"):
    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')

    # 發送HTML格式郵件
    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 為 SMTP 端口號
        # 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("郵件發送成功")
    except Exception as e:
        print(e)
    # except smtplib.SMTPException:
        # print("Error: 無法發送郵件")

3.主要功能代碼

收集到的DB數據文件的信息保存到表mssql_dblogsize中，其建表的腳本如下：

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

為瞭方便對表mssql_dblogsize的數據進行管理和展示，在其基礎上抽象加工出瞭一個視圖v_mssql_dblogsize，註意Size大小的轉換（Size/128/1024 as SizeGB）

創建視圖的腳本如下：

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

本測試實例使用的數據庫為qqDB，監控的各個DB Server保存在瞭表QQDBServer中，註意Port 不一定為標準端口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')

    # 獲取需要遍歷的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()

告警郵件的功能實現為mssql_alert_dblogsize.py,此份代碼的告警閾值設置的為50G，數據來自於視圖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 = "[email protected]"


    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')

    # 獲取需要遍歷的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>" + "以下數據庫的db log文件，已大於50G.請及時檢查，謝謝！ " + "<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.實現

定時任務是通過windows的計劃任務來實現的,在此不做過多的敘述。告警郵件的部分截圖如下:

5.附錄

1.報錯定位，判斷是不是log文件過大

https://blog.csdn.net/weixin_30785593/article/details/99912405

2.關於為什麼數據庫log文件過大，我們可以參考以下分享的文章

https://blog.csdn.net/chinadm123/article/details/44941275

到此這篇關於通過Python實現對SQL Server 數據文件大小的監控告警的文章就介紹到這瞭,更多相關PythonSQL Server 數據監控告警內容請搜索WalkonNet以前的文章或繼續瀏覽下面的相關文章希望大傢以後多多支持WalkonNet！

推薦閱讀：

    None Found