SQL SERVER占用CPU过高排查和优化

转载自：https://www.cnblogs.com/vice/p/9163241.html

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

1 use master
2 go
3 --如果要指定数据库就把注释去掉
4 select * from sys.[sysprocesses] where [spid]>50 --and db_name([dbid])='gposdb'

　　

　　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。

按 ctrl+c 复制代码

按 ctrl+c 复制代码

查看是哪些sql语句占用较大可以使用下面代码

按 ctrl+c 复制代码

按 ctrl+c 复制代码

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

 1 select top 10
 2  [session_id],
 3  [request_id],
 4  [start_time] as '开始时间',
 5  [status] as '状态',
 6  [command] as '命令',
 7  dest.[text] as 'sql语句', 
 8  db_name([database_id]) as '数据库名',
 9  [blocking_session_id] as '正在阻塞其他会话的会话id',
10  der.[wait_type] as '等待资源类型',
11  [wait_time] as '等待时间',
12  [wait_resource] as '等待的资源',
13  [dows].[waiting_tasks_count] as '当前正在进行等待的任务数',
14  [reads] as '物理读次数',
15  [writes] as '写次数',
16  [logical_reads] as '逻辑读次数',
17  [row_count] as '返回结果行数'
18  from sys.[dm_exec_requests] as der 
19  inner join [sys].[dm_os_wait_stats] as dows 
20  on der.[wait_type]=[dows].[wait_type]
21  cross apply 
22  sys.[dm_exec_sql_text](der.[sql_handle]) as dest 
23  where [session_id]>50  
24  order by [cpu_time] desc

4、查询cpu占用最高的sql语句

 1 select top 10
 2    total_worker_time/execution_count as avg_cpu_cost, plan_handle,
 3    execution_count,
 4    (select substring(text, statement_start_offset/2 + 1,
 5       (case when statement_end_offset = -1
 6          then len(convert(nvarchar(max), text)) * 2
 7          else statement_end_offset
 8       end - statement_start_offset)/2)
 9    from sys.dm_exec_sql_text(sql_handle)) as query_text
10 from sys.dm_exec_query_stats
11 order by [avg_cpu_cost] desc

5、索引缺失查询

 1 select 
 2     databasename = db_name(database_id)
 3     ,[number indexes missing] = count(*) 
 4 from sys.dm_db_missing_index_details
 5 group by db_name(database_id)
 6 order by 2 desc;
 7 select  top 10 
 8         [total cost]  = round(avg_total_user_cost * avg_user_impact * (user_seeks + user_scans),0) 
 9         , avg_user_impact
10         , tablename = statement
11         , [equalityusage] = equality_columns 
12         , [inequalityusage] = inequality_columns
13         , [include cloumns] = included_columns
14 from        sys.dm_db_missing_index_groups g 
15 inner join    sys.dm_db_missing_index_group_stats s 
16        on s.group_handle = g.index_group_handle 
17 inner join    sys.dm_db_missing_index_details d 
18        on d.index_handle = g.index_handle
19 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。