当前位置:网站首页>Dameng AWR report extraction
Dameng AWR report extraction
2022-07-21 03:26:00 【qq_ thirty-eight million one hundred and thirteen thousand four】
Access method :
Call the system procedure and set the snapshot interval :
SP_INIT_AWR_SYS(1);
— Set the snapshot interval to 10 minute
CALL DBMS_WORKLOAD_REPOSITORY.AWR_SET_INTERVAL(10);
Create a snapshot manually :
DBMS_WORKLOAD_REPOSITORY.CREATE_SNAPSHOT();
View snapshot information :
SELECT * FROM SYS.WRM$_SNAPSHOT;
Generate snapshot Of id stay 1~2 Within the scope of AWR The report :
SELECT * FROM TABLE (DBMS_WORKLOAD_REPOSITORY.AWR_REPORT_HTML(1,2));
— Save to /home/dmdba In the catalog dm7_awr.html In file
SYS.AWR_REPORT_HTML(1,2,’/home/dmdba’,’dm7_awr.html’);
AWR Report daily analysis
3.1 SQL ordered by Elapsed Time
The total execution time is recorded TOP SQL( Please note that it is within the scope of monitoring that SQL Total execution time of , Not a single time SQL execution time Elapsed Time = CPU Time + Wait Time).
Elapsed Time(S): SQL Total time taken for statement execution , This sort is done according to this field . Note that this time is not a single SQL Running time , It's within the scope of monitoring SQL The sum of execution times . The unit time is seconds .Elapsed Time = CPU Time + Wait Time
CPU Time(s): by SQL Statement execution time CPU Total occupation time , This time will be less than or equal to Elapsed Time Time . The unit time is seconds .
Executions: SQL The total number of executions of the statement within the monitoring range .
Elap per Exec(s): Do it once SQL The average time . The unit time is seconds .
% Total DB Time: by SQL Of Elapsed Time Time as a percentage of total database time .
SQL ID: SQL Of the statement ID Number , After clicking, you can navigate to the... Below SQL In the detailed list , Click on IE The return of can return to the current SQL ID The place of .
SQL Module: According to the SQL How to connect to the database for execution , If it is to use SQL*Plus perhaps PL/SQL The link is basically someone debugging the program . It is usually executed by linking the foreground application sql The location is empty .
SQL Text: ordinary sql Tips , Click... For details SQL ID.
3.2 SQL ordered by CPU Time:
Record the execution of CPU The sum of time is the longest TOP SQL( Please note that it is within the scope of monitoring that SQL The execution accounts for CPU The sum of time , Not a single time SQL execution time ).
3.3 SQL ordered by Gets:
Record the execution of the total buffer gets( Logic IO) Of TOP SQL( Please note that it is within the scope of monitoring that SQL The execution accounts for Gets The sum of the , Not a single time SQL Execution takes up Gets).
3.4 SQL ordered by Reads:
Records the total disk physical reads performed ( Physics IO) Of TOP SQL( Please note that it is within the scope of monitoring that SQL The execution of accounts for the total physical reads on the disk , Not a single time SQL Physical read on disk occupied by execution ).
3.5 SQL ordered by Executions:
Recorded according to SQL The number of times of execution ordered TOP SQL. In this sort, we can see SQL Number of executions .
3.6 SQL ordered by Parse Calls:
Recorded SQL Of the number of soft resolution TOP SQL. Speaking of soft parsing (soft prase) And hard parsing (hard prase) see :http://www.taoyoyo.net/ttt/post/305.html
3.7 SQL ordered by Sharable Memory:
Recorded SQL Occupy library cache The size of the TOP SQL.Sharable Mem (b): Occupy library cache Size , The unit is byte.
3.8 SQL ordered by Version Count:
Recorded SQL To open a child cursor TOP SQL.
3.9 SQL ordered by Cluster Wait Time:
It records the waiting time of the cluster TOP SQL
https://eco.dameng.com/
边栏推荐
- Better to struggle for the future than regret the past
- 输入一个整数和一棵二元树
- antd mobile 表单验证 rc-form 使用
- (四)PyTorch深度学习:PytTorch实现线性回归
- 一致性哈希,虚拟节点,布隆过滤器
- Spark SQL 中的小 TIPS
- [dish of learning notes, dog learning C] first learn operators and original code, inverse code, complement code
- mysql_备份还原_指定表_备份表_还原表_innobackup
- 事务(隔离界别)
- 分别用递归和非递归的方式实现二叉树先序、中序和后序遍历
猜你喜欢
Chrome 进程架构
Knapsack problem (01 knapsack / full knapsack explanation)
(五)PyTorch深度学习:Logistic回归
根据不同configuration加载不同的plist文件
Configuration of Visual Studio development environment
Hyperledger Fabric 超级账本 CA搭建使用
Centos8 (Linux) 安装 redis
小白教程--anaconda之Jupyter Notebook自动补全配置教程
(五)PyQt5系列教程:使用Pycharm对PyQt5在串口助手参数选项进行内部逻辑设计(二)
Redis persistence
随机推荐
阿里矢量图库 当前页全选
手写Promise
(八)PyTorch深度学习:卷积神经网络(基础)---将(七)全连接神经网络改成卷积神经网络
kettle_配置数据库连接_报错
达梦ODBC安装
列表元素相加
SQL事务
达梦统计信息收集
打印两个有序链表的公共部分
[dish of learning notes dog learning C] detailed operator
【深度学习】怎么为模型选择正确的最后一层激活和损失函数
数组常用方法
SQL处理数据 同期群分析
Los Angeles: t226229 arithmetic series
小白教程--anaconda之Jupyter Notebook自动补全配置教程
用户登录Demo
ECMAScript新特性
(五)PyTorch深度学习:Logistic回归
Redis持久化
Configuration of Visual Studio development environment