tag:blogger.com,1999:blog-1550840387946350137.post6660926247426426841..comments2022-04-03T00:39:35.304-07:00Comments on Performance Engineering and Capacity Planning: Understanding CPU Time as an Oracle Wait eventUnknownnoreply@blogger.comBlogger5125tag:blogger.com,1999:blog-1550840387946350137.post-29216027774543642722011-06-02T10:20:13.280-07:002011-06-02T10:20:13.280-07:00For monitoring CPU usage, I guess you could use v$...For monitoring CPU usage, I guess you could use v$osstat. As always, you would need to correlate using OS tools such as vmstat or mpstat. <br /><br />The v$sys_time_model may be a better view to see if you have a CPU bottleneck. As it is cumulative, you would need to take multiple snapshots over a period of time.SSKhttps://www.blogger.com/profile/01063233277470336458noreply@blogger.comtag:blogger.com,1999:blog-1550840387946350137.post-78327500180185672852011-06-01T20:15:47.011-07:002011-06-01T20:15:47.011-07:00Krishna,
Thanks for the article,nice.
As a DBA...Krishna, <br /><br />Thanks for the article,nice. <br /><br />As a DBA I want to limit myself to database front and let the SA figureout what's wrong with CPU if I confirm that nothing is wrong from Oracle Database prespecitive . What in your view is the best way to monitor CPU usage within oracle Database. <br />As you said , CPU as top event (time) is a indication that something is consuming more resource. I would like to quickly run some queries against v$ views to get the information about the same . <br />Do you thing V$OSSTAT or V$SYSSTAT are good starting point ? what should be the chronological steps to come to a conclude the bottleneck ?UKnoreply@blogger.comtag:blogger.com,1999:blog-1550840387946350137.post-68337339791863539842010-10-05T05:42:14.030-07:002010-10-05T05:42:14.030-07:00Excellent topic will use some of the tips
Capacity...Excellent topic will use some of the tips<br /><a href="http://www.optisolutionsus.com" rel="nofollow">Capacity Planning</a><br /><a href="http://www.optisolutions.co.za" rel="nofollow">Capacity Planning</a>Sharee-Leenoreply@blogger.comtag:blogger.com,1999:blog-1550840387946350137.post-85652206901330951132010-02-06T18:26:52.179-08:002010-02-06T18:26:52.179-08:00"Besides fixed time another reason why a proc..."Besides fixed time another reason why a process would be switched out of a processor would be when it makes an I/O call. The process would then relinquish the CPU. Would changing the "fixed quanta" affect this behaviour ?"<br /><br />I do not believe so. The time quanta would apply only when the process is running on the CPU actively. <br /><br />The correlation of priorities with the numbers as displayed by the "ps" command depends on the argument passed to "ps".<br /><br />With the -c flag, higher numbers in the PRI column means higher priority. Otherwise, it would mean the opposite - lower numbers are higher priority.SSKhttps://www.blogger.com/profile/01063233277470336458noreply@blogger.comtag:blogger.com,1999:blog-1550840387946350137.post-55601031648876703162010-02-06T08:33:28.494-08:002010-02-06T08:33:28.494-08:00"it has the potential to completely break the..."it has the potential to completely break the environment if not done correctly." is very significant.<br /><br />Besides fixed time another reason why a process would be switched out of a processor would be when it makes an I/O call. The process would then relinquish the CPU. Would changing the "fixed quanta" affect this behaviour ?<br /><br />(Also, I need to recheck how a Priority is listed. I thought a higher absolute value for PRI meant a lower relative priority).Hemant K Chitalehttps://www.blogger.com/profile/07369112096230549250noreply@blogger.com