Activating a Performance Trace
来源:互联网 发布:软件开发属于什么行业 编辑:程序博客网 时间:2024/06/05 09:00
Use a different name if you want to trace the actions of another user. The user who activates the trace does not have to be the same one whose actions are being traced.
Information on activating a trace
Keep in mind the following points when activating a trace:
->Ensure that the user whose actions are to be recorded carries out only one action during the trace; otherwise, the trace will not be clear. You should also ensure that no background jobs or update requests are running for this user.
->The performance trace is activated in the application server. For each database operation, data is written to a trace file in the file system on the application server. You must therefore ensure that you have logged on to the same application server as the user to be monitored. This is particularly important if you want to record an update request or background job and are working in a system with distributed updating or distributed background processing. In this case, you will not know where the request will be started and, as a result, will have to start the trace on all application servers with update or background work processes.
->The SQL trace displays only accesses to the database. SQL statements that can be satisfied from datain the SAP buffer do not appear in the trace. If you want to analyze SAP buffer accesses, activate the SAP buffer trace.
->However, buffer load processes are also recorded in the SQL trace. Since you are normally not interested in recording the buffer load process in the SQL trace, first execute a program once without activating the trace to allow the buffers to be loaded (that is, the SAP buffers and database buffers). Then, run the program again with the SQL trace activated and use the results of this trace for evaluation .
->During the trace, look at the following monitors: the work process overview (for general monitoring), the operating system monitor of the database server (for monitoring possible CPU bottlenecks on the database server), and the database process monitor for direct monitoring of the executed SQL statements. It makes no sense to watch these monitors during the trace if you are logged on as the user being traced. The SQL statements of the monitors would appear in the trace and thus make the trace unreadable .
->The default trace file name is set with the SAP profile parameter rstr/file. In the initial screen, you can assign a different name to the trace
file. Writing to the trace file is cyclical in the sense that, when the file is full, the oldest entries are deleted to make room for new entries. The size of the trace file (in bytes) is specified by the SAP profile parameterrstr/max_diskspace, for which the default value is 16.384,000bytes (16MB) .
->You can also store recorded traces. For this purpose, select PERFORMANCE TRACE • SAVE TRACE or PERFORMANCE TRACE • DISPLAY SAVED TRACE to retrieve a saved trace again.
- Activating a Performance Trace
- Performance Trace
- Performance Trace
- How to run ST12 to do a performance trace
- kvm trace performance events
- Could not open performance trace file
- android performance trace使能cpufreq_gov
- 6: Activating A Virtualenv( Command line Python scripting )
- Getting a Packet Trace
- A Taxonomy for Performance
- How To Collect 10046 Trace (SQL_TRACE) Diagnostics for Performance Issues
- Access denied when deploying a timer Job or activating a feature
- A Class for Creating a Trace Log
- Scheduling a SQL Server Profiler Trace
- gstack - stack trace of a running process
- TRACE
- trace
- TRACE
- ajax02基于xml的实现
- 3D语音天气球(源码分享)——完结篇
- Cocos Studio和Cocos2d-x版本对应关系
- 各种排序算法及其java程序实现
- tomcat启动解决端口占用问题
- Activating a Performance Trace
- 设置session时长,超时间需重新登录
- (hibernate之三)session的delete方法(转的学习用)
- string类的构造函数、析构函数、拷贝构造函数和赋值函数
- WebRtc 的VAD算法解析
- Unity使用OpenCvSharp进行人脸识别
- linux shell随机数获取的方法
- iOS开发ASIHTTPRequest身份验证【5】
- std::vector的reserve和resize的区别