本篇博客总结一下Python采集SQL Server数据库服务器的磁盘使用信息，其实这里也是根据需求不断推进演化的一个历程，我们监控服务器的磁盘走了大概这样一个历程：
1：使用SQL Server作业监控磁盘空间
很久之前写过一篇博客“MS SQL 监控磁盘空间告警”，后面对这个脚本进行过多次完善和优化，做成了一个模板。在每台SQL Server服务器上都部署了，确实也很实用。告警也很给力，但是缺点也非常明显。
优点：1： 自己动手DIY，在没有部署运维工具的前提下，确实能提前预警，抛开不足来说，告警还是非常给力的。
缺点: 1： 每台服务器都需要部署，升级也很是麻烦。
2： 数据分散，在模式上有致命的先天不足。监控工具一般为星型结构，采集集中数据。
每台服务器都需要部署，如果有修改，每台服务器都需要发布更新，维护管理不方便。采集的一些数据分散，每台SQL Server数据库都需要保存一点数据。
3： 通用性差，只能监控SQL Server服务器，Linux服务器，我们用的是crontab跑shell+perl脚本来监控磁盘空间并告警。
2： Zabbix监控磁盘告警
后面部署了Zabbix监控工具，Zabbix监控工具功能强大，不仅仅提供了磁盘空间告警功能，还提供了监控磁盘I/O等功能。更重要的是通用性很强大：只要是服务器就能监控，而方法1仅仅能监控SQL Server数据库服务器。
优点： 1： Zabbix监控工具功能强大
2： Zabbix监控工具通用性强
缺点： 1：需要分析磁盘空间的历史数据比较麻烦，二次开发也比较麻烦（个人对zabbix了解不深入，可能对于高手而言，也是非常容易简单的事情，仅仅是个人的一点体会感受）。例如我要获取某个服务器的历史数据，对磁盘的增长情况做分析。需要关联好多表，非常麻烦。简单研究后就直接放弃了。
自从Zabbix监控工具上线后，方法1就显得可有可无。基本上处于被替换的尴尬境地。
3：使用Python脚本采集
这里存粹是为了扩展我自己的工具MyDB的功能，顺手写点Python脚本练手，而且目前而言，只能采集SQL Server服务器的磁盘空间使用情况。功能和通用性不能和Zabbix监控工具比。功能有很多局限性和不足之处，通用性也很差，但是也有一些不错的优点。
优点： 1：不需要部署客户端，也不需要每台服务器去部署（与方法1对比而言）。
2：批量采集，集中保存采集数据。方便统一告警，数据分析。
3：简洁与简单，灵活性高：目前就2个表，一个表[dbo].[SERVER_DISK_INFO]保存最近一次采集的磁盘空间使用情况数据，另外一个表[dbo].[SERVER_DISK_INFO_HIS]保存历史数据，可以做一些扩容分析等。
例如，磁盘空间告警了，系统管理员会咨询我，如果进行扩容，需要多大的空间，保证半年内，不会再次出现告警，那么就可以一个脚本计算一下（平均每个月增长值* 月数）
缺点： 1：通用性差，目前只能采集SQL Server数据库服务器的磁盘信息。后续再考虑扩展性。实在没有精力一步到位，慢慢完善扩充。
2：功能单一，不像Zabbix，还可以监控磁盘I/O,这个Python脚本就只能采集磁盘空间使用率，并不能扩展其功能，有一定局限性。
脚本get_win_disk_info.py如下所示：
# -*- coding: utf-8 -*-'''--------------------------------------------------------------------------------------------- Script Name : get_win_disk_info.py-- Script Auotor : 潇湘隐者-- Script Description : 采集SQL Server数据库的磁盘使用数据，方便统一分析和告警处理！-------------------------------------------------------------------------------------------'''import pymssqlimport loggingimport os.pathimport base64from cryptography.fernet import Fernet
# 第一步，创建一个loggerlogger = logging.getLogger()logger.setLevel(logging.DEBUG) # Log等级开关# 第二步，创建一个handler，用于写入日志文件#log_path = os.path.dirname(os.getcwd()) + '/logs/'log_path = '/home/konglb/logs/'log_name = log_path + 'get_win_disk_info.log'logfile = log_namefile_handler = logging.FileHandler(logfile, mode='a+')file_handler.setLevel(logging.ERROR) # 输出到file的log等级的开关# 第三步，定义handler的输出格式formatter = logging.Formatter("%(asctime)s - %(filename)s[line:%(lineno)d] - %(levelname)s: %(message)s")file_handler.setFormatter(formatter)# 第四步，将handler添加到logger里面logger.addHandler(file_handler)# 如果需要同時需要在終端上輸出，定義一個streamHandlerprint_handler = logging.StreamHandler() # 往屏幕上输出print_handler.setFormatter(formatter) # 设置屏幕上显示的格式logger.addHandler(print_handler)
key=bytes(os.environ.get('key'),encoding="utf8")
cipher_suite = Fernet(key)with open('/home/konglb/python/conf/ms_db_conf.bin', 'rb') as file_object: for line in file_object: encryptedpwd = linedecrypt_pwd = (cipher_suite.decrypt(encryptedpwd))password_decrypted = bytes(decrypt_pwd).decode("utf-8") #convert to stringenv_db_user=os.environ.get('db_user')db_user=base64.b64decode(bytes(env_db_user, encoding="utf8"))
try: dest_db_conn = pymssql.connect(host=os.environ.get('db_host'), user=bytes.decode(db_user), password=password_decrypted, database='DATABASE_REPOSITORY', charset="utf8");key=bytes(os.environ.get('key'),encoding="utf8")
# cursor = dest_db_conn.cursor(); # as_dict(bool) ：如果设置为True，则后面的查询结果返回的是字典，关键字为查询结果的列名；否则(默认)返回的为list。 # 可以通过在创建游标时指定as_dict参数来使游标返回字典变量,字典中的键为数据表的列名 cursor = dest_db_conn.cursor(as_dict=True) #DELETE FROM [dbo].[DB_JOB_RUN_ERROR] # WHERE RUN_DATE_TIME >= CAST(CONVERT(VARCHAR(10),GETDATE(),120) AS DATETIME); sql_text = """INSERT INTO dbo.SERVER_DISK_INFO_HIS ( COLLECT_DATE , FACTORY_CD , SERVER_NAME , DISK_NAME , TOTAL_SPACE , USED_SPACE , FREE_SPACE , FREE_PERCENT ) SELECT COLLECT_DATE , FACTORY_CD , SERVER_NAME , DISK_NAME , TOTAL_SPACE , USED_SPACE , FREE_SPACE , FREE_PERCENT FROM [dbo].[SERVER_DISK_INFO]; TRUNCATE TABLE [dbo].[SERVER_DISK_INFO]; """ cursor.execute(sql_text); dest_db_conn.commit()
sql_text = """ SELECT SERVER_CD , SERVER_IP , USER_NAME , dbo.DecryptByPassPhrasePwd(PASSWORD) AS PASSWORD , SERVER_NAME, DB_VERSION , INSTANCE_NAME FROM dbo.DB_SERVER_CONFIG WHERE DATABASE_TYPE = 'SQL SERVER' AND COLLECT_DATA = 1; """
cursor.execute(sql_text); rows = cursor.fetchall();
for row in rows:
try: src_db_conn = pymssql.connect(host=row['SERVER_IP'], user=row['USER_NAME'], password=row['PASSWORD'], database='master', charset="utf8", autocommit=True);
sub_cursor = src_db_conn.cursor(as_dict=True)
if row['DB_VERSION'] <= 2000: logger.info(row['SERVER_NAME'] + ' not gather') continue else: #logger.info(row['DB_VERSION']) sql_db_patch="SELECT SERVERPROPERTY('productlevel') AS PRODUCT_LEVEL" sub_cursor.execute(sql_db_patch) db_patch = sub_cursor.fetchone() #必须转换，否则返回的为bytes，不是str patch_info= str(db_patch['PRODUCT_LEVEL'], encoding = "utf-8")
if ((row['DB_VERSION']== 2005) or (row['DB_VERSION'] ==2008 and patch_info == 'RTM')): #logger.info(row['SERVER_NAME'] + ' patch is ' + patch_info) #continue
sql_job_info=""" DECLARE @Result INT; DECLARE @objectInfo INT; DECLARE @DriveInfo CHAR(1); DECLARE @TotalSize VARCHAR(20); DECLARE @OutDrive INT; DECLARE @UnitGB FLOAT; DECLARE @CurrentDate DATETIME; DECLARE @ConfValue INT; SET @UnitGB = 1073741824.0; --创建临时表保存服务器磁盘容量信息 CREATE TABLE #DiskCapacity ( COLLECT_DATE DATETIME , FACTORY_CD NVARCHAR(24), SERVER_NAME NVARCHAR(64), DISK_NAME NVARCHAR(2) , TOTAL_SPACE FLOAT, USED_SPACE FLOAT, FREE_SPACE FLOAT, FREE_PERCENT FLOAT ); INSERT #DiskCapacity (DISK_NAME,FREE_SPACE ) EXEC master.dbo.xp_fixeddrives; EXEC sp_configure 'show advanced options', 1 RECONFIGURE WITH OVERRIDE; SELECT @ConfValue = value FROM sys.sysconfigures WHERE comment LIKE '%Ole Automation Procedures%' IF @ConfValue = 0 BEGIN EXEC sp_configure 'Ole Automation Procedures', 1; RECONFIGURE WITH OVERRIDE; END EXEC @Result = master.sys.sp_OACreate 'Scripting.FileSystemObject',@objectInfo OUT; DECLARE CR_DiskInfo CURSOR LOCAL FAST_FORWARD FOR SELECT DISK_NAME FROM #DiskCapacity ORDER by DISK_NAME OPEN CR_DiskInfo; SET @CurrentDate = GETDATE(); FETCH NEXT FROM CR_DiskInfo INTO @DriveInfo WHILE @@FETCH_STATUS=0 BEGIN EXEC @Result = sp_OAMethod @objectInfo,'GetDrive', @OutDrive OUT, @DriveInfo EXEC @Result = sp_OAGetProperty @OutDrive,'TotalSize', @TotalSize OUT UPDATE #DiskCapacity SET TOTAL_SPACE=ROUND(@TotalSize/@UnitGB,2), COLLECT_DATE=@CurrentDate, FACTORY_CD=%s, SERVER_NAME=@@SERVERNAME, --USED_SPACE=(@TotalSize-FREE_SPACE)/@UnitGB, --FREE_PERCENT=FREE_SPACE/@TotalSize*100, FREE_SPACE=ROUND(FREE_SPACE/1024,2) WHERE DISK_NAME =@DriveInfo UPDATE #DiskCapacity SET USED_SPACE=(TOTAL_SPACE-FREE_SPACE), FREE_PERCENT=ROUND(FREE_SPACE/TOTAL_SPACE*100,2) WHERE DISK_NAME =@DriveInfo FETCH NEXT FROM CR_DiskInfo INTO @DriveInfo END CLOSE CR_DiskInfo DEALLOCATE CR_DiskInfo; EXEC @Result=sp_OADestroy @objectInfo EXEC sp_configure 'show advanced options', 1 RECONFIGURE WITH OVERRIDE; IF @ConfValue = 0 BEGIN EXEC sp_configure 'Ole Automation Procedures', 0; RECONFIGURE WITH OVERRIDE; END EXEC sp_configure 'show advanced options', 0 RECONFIGURE WITH OVERRIDE; SELECT * FROM #DiskCapacity """ sub_cursor.execute(sql_job_info, row['SERVER_CD']); job_rows = sub_cursor.fetchall(); src_db_conn.close() error_job_info = [] for sub_row in job_rows: data = ( sub_row['COLLECT_DATE'], sub_row['FACTORY_CD'], sub_row['SERVER_NAME'], sub_row['DISK_NAME'], sub_row['TOTAL_SPACE'], sub_row['USED_SPACE'], sub_row['FREE_SPACE'], sub_row['FREE_PERCENT']) error_job_info.append(data)
''' logger.info('2008 2005') logger.info(row['SERVER_NAME']) sub_cursor.callproc('[msdb].[dbo].[sp_get_diskinfo]') result_rows =sub_cursor.fetchall() src_db_conn.close() error_job_info = [] for sub_row in result_rows: data = ( sub_row['COLLECT_DATE'], sub_row['FACTORY_CD'], sub_row['SERVER_NAME'], sub_row['DISK_NAME'], sub_row['TOTAL_SPACE'], sub_row['USED_SPACE'], sub_row['FREE_SPACE'], sub_row['FREE_PERCENT']) error_job_info.append(data) ''' else:
sql_job_info = """WITH Server_Disk AS ( SELECT DISTINCT REPLACE(vs.volume_mount_point, ':\\' , '') AS DISK_NAME, CAST(VS.total_bytes/1024.0/1024/1024 AS NUMERIC(18,2) ) AS [TOTAL_SPACE], CAST(VS.available_bytes/1024.0/1024/1024 AS NUMERIC(18,2)) AS [FREE_SPACE] FROM sys.master_files AS f CROSS APPLY sys.dm_os_volume_stats(f.database_id, f.file_id) AS vs ) SELECT GETDATE() AS COLLECT_DATE, %s AS FACTORY_CD , @@SERVERNAME AS SERVER_NAME , D.DISK_NAME AS DISK_NAME, D.[TOTAL_SPACE] AS TOTAL_SPACE, D.[TOTAL_SPACE] - D.[FREE_SPACE] AS USED_SPACE, D.[FREE_SPACE] AS FREE_SPACE, CAST(D.[FREE_SPACE] * 100 / D.[TOTAL_SPACE] AS NUMERIC(18, 2)) AS FREE_PERCENT FROM Server_Disk AS D ORDER BY D.DISK_NAME; """
sub_cursor.execute(sql_job_info, row['SERVER_CD']); job_rows = sub_cursor.fetchall(); src_db_conn.close() error_job_info = [] for sub_row in job_rows: data = (sub_row['COLLECT_DATE'], sub_row['FACTORY_CD'], sub_row['SERVER_NAME'], sub_row['DISK_NAME'], sub_row['TOTAL_SPACE'], sub_row['USED_SPACE'], sub_row['FREE_SPACE'], sub_row['FREE_PERCENT']) error_job_info.append(data)
save_job_info = """ INSERT INTO dbo.SERVER_DISK_INFO ( COLLECT_DATE , FACTORY_CD , SERVER_NAME , DISK_NAME , TOTAL_SPACE , USED_SPACE , FREE_SPACE , FREE_PERCENT ) VALUES(%s,%s,%s,%s,%d,%d,%d,%d)""" cursor.executemany(save_job_info, error_job_info); dest_db_conn.commit() logger.info(row['SERVER_NAME'] + ' gather successful')
except pymssql.InterfaceError as fe: logger.error(fe.message) except pymssql.DatabaseError as e: dest_db_conn.rollback(); logger.error(row['SERVER_IP'] + ' 采集出错，请检查处理异常') logger.error(e) finally: src_db_conn.close()except pymssql.InterfaceError as fe: logger.error(fe.message)except pymssql.DatabaseError as e: dest_db_conn.rollback();
logger.error(row['SERVER_IP'] + ' 采集出错，请检查处理异常') logger.error(e)finally: dest_db_conn.close()
