I was thinking that generating and understanding the AWR report is very difficult and complex. But, this video made me very clear in one shot. I have a request for you. Kindly upload videos for performance tuning. I got stuck in every interview when they ask about the performance tuning. Can you please guide step step from where to start the tuning of an existing code or sql. The various techniques that can be applied to solve a problem. Thank you in advance.
Are you comparing my WINDOWS machine time with AWR time? The AWR report will be generated as per the Linux server time. Probably Linux machine time might be different, in my case.
Hi Arun: I have a Question! Like I got a perf issue where a migration script command was run in Dev env and it ran for almost 1 hrs and completed successfully. Now the same command was run on Staging (with the same infra configuration and indexes), it got stuck and took more than 10+ hrs and isn't getting completed. Checking Locks, ASH, and AWR didn't find a solution. What could be the possibilities for this issue?
My tomcat server is slow since few days now. I have the awr report.. i am just a beginner can you please help me. Which part of report i should take a look at in order to find root cause of slowness
when talking about increasing instance(as you check whether or not the instance needs to be tunned)....what specific parameter are you referring to? Thank you.
Instance is nothing but sga +bg process so there should be sga size to increase or decrese depentents on the statistics of awr analysis he mean to say i guess
Hi, very clean explanation. I have a question, if in Top 10 Foreground Events by Total Wait Time, my first event is DB CPU with %DB TIME 69.2 and total wait time (sec) is 55.3k, should I consider a problem? In the second event is latch free with %DB Time 7.7 and total wait time (sec) is 6187.6. What should I do with latch free, please?
Hey can you explain a bit about DB CPU wait event and what to do to fix it. It is coming at the top of the Top 10 Foreground Events by Total Wait Time with 46.6K Total Wait Time (sec)
In the video, at times 56:00 and 1:05:00, your interpretation of %CPU is incorrect and misguiding. You are meaning, %CPU of 95 is a problem, which is not correct. Ideally a good query should have %CPU as 100. Queries running without any waits will have %CPU as 100 which is perfect from the database instance configuration. Usually queries blocked due to "row lock contention" will have the %CPU as almost zero which is a problem.
You should have had a real report, atleast from Perf env, and explained it, instead of showing a dataless report. I am grateful for the series but hearing that guy asking you for more info and you shutting him down makes me wonder, you could have explained it simply in the time you spent convincing him.
Thanks Arun, appreciate your effort to explain AWR Report, It was very helpful for me for job responsibility, but at the same time expectation is more, pls create more session on AWR and most optimize oracle database configuration.
