SQL SERVER占用CPU过高优化

小木木

　　操作系统是Windows2008R2 ,数据库是SQL2014 64位。
　　近阶段服务器出现过几次死机，管理员反馈机器内存使用率100%导致机器卡死。于是做了个监测服务器的软件实时记录CPU数据，几日观察得出数据如下：

　　SQL优化方法：
　　1、查看连接对象



USE master
GO
SELECT * FROM sys.[sysprocesses] WHERE [spid]>50

　　当前连接对象有67个其中‘WINAME’的主机名，‘jTDS’的进程名不属于已知常用软件，找到这台主机并解决连接问题。在360流量防火墙中查看有哪个软件连接了服务器IP，除之。

　　2、然后使用下面语句看一下各项指标是否正常，是否有阻塞，正常情况下搜索结果应该为空。



SELECT TOP 10
[session_id],
[request_id],
[start_time] AS '开始时间',
[status] AS '状态',
[command] AS '命令',
dest.[text] AS 'sql语句',
DB_NAME([database_id]) AS '数据库名',
[blocking_session_id] AS '正在阻塞其他会话的会话ID',
[wait_type] AS '等待资源类型',
[wait_time] AS '等待时间',
[wait_resource] AS '等待的资源',
[reads] AS '物理读次数',
[writes] AS '写次数',
[logical_reads] AS '逻辑读次数',
[row_count] AS '返回结果行数'
FROM sys.[dm_exec_requests] AS der
CROSS APPLY
sys.[dm_exec_sql_text](der.[sql_handle]) AS dest
WHERE [session_id]>50 AND DB_NAME(der.[database_id])='gposdb'  
ORDER BY [cpu_time] DESC
　　查看是哪些SQL语句占用较大可以使用下面代码




SELECT TOP 10
dest.[text] AS 'sql语句'
FROM sys.[dm_exec_requests] AS der
CROSS APPLY
sys.[dm_exec_sql_text](der.[sql_handle]) AS dest
WHERE [session_id]>50  
ORDER BY [cpu_time] DESC
　　3、如果SQLSERVER存在要等待的资源，那么执行下面语句就会显示出会话中有多少个worker在等待



SELECT TOP 10
  [session_id],
  [request_id],
  [start_time] AS '开始时间',
  [status] AS '状态',
  [command] AS '命令',
  dest.[text] AS 'sql语句',
  DB_NAME([database_id]) AS '数据库名',
  [blocking_session_id] AS '正在阻塞其他会话的会话ID',
  der.[wait_type] AS '等待资源类型',
  [wait_time] AS '等待时间',
  [wait_resource] AS '等待的资源',
  [dows].[waiting_tasks_count] AS '当前正在进行等待的任务数',
  [reads] AS '物理读次数',
  [writes] AS '写次数',
  [logical_reads] AS '逻辑读次数',
  [row_count] AS '返回结果行数'
  FROM sys.[dm_exec_requests] AS der
  INNER JOIN [sys].[dm_os_wait_stats] AS dows
  ON der.[wait_type]=[dows].[wait_type]
  CROSS APPLY
  sys.[dm_exec_sql_text](der.[sql_handle]) AS dest
  WHERE [session_id]>50  
  ORDER BY [cpu_time] DESC
　　4、查询CPU占用最高的SQL语句



SELECT TOP 10
    total_worker_time/execution_count AS avg_cpu_cost, plan_handle,
    execution_count,
    (SELECT SUBSTRING(text, statement_start_offset/2 + 1,
       (CASE WHEN statement_end_offset = -1
          THEN LEN(CONVERT(nvarchar(max), text)) * 2
          ELSE statement_end_offset
       END - statement_start_offset)/2)
    FROM sys.dm_exec_sql_text(sql_handle)) AS query_text
FROM sys.dm_exec_query_stats
ORDER BY [avg_cpu_cost] DESC
　　5、索引缺失查询



SELECT
     DatabaseName = DB_NAME(database_id)
     ,[Number Indexes Missing] = count(*)
FROM sys.dm_db_missing_index_details
GROUP BY DB_NAME(database_id)
ORDER BY 2 DESC;
SELECT  TOP 10
         [Total Cost]  = ROUND(avg_total_user_cost * avg_user_impact * (user_seeks + user_scans),0)
         , avg_user_impact
         , TableName = statement
         , [EqualityUsage] = equality_columns
         , [InequalityUsage] = inequality_columns
         , [Include Cloumns] = included_columns
FROM        sys.dm_db_missing_index_groups g
INNER JOIN    sys.dm_db_missing_index_group_stats s
        ON s.group_handle = g.index_group_handle
INNER JOIN    sys.dm_db_missing_index_details d
        ON d.index_handle = g.index_handle
ORDER BY [Total Cost] DESC;
　　找到索引缺失的表，根据查询结果中的关键次逐一建立索引。
　　做完这些测试，基本能找到问题。