10-31-2019 4:35 PM
Hi ,
We are running a program in background in pre-production system(Data for this system is refreshed just last week) and i is taking very long.When i look in SM50 i can see a work process stuck 'Sequential Read of DB DFKKOP'.
But the same program runs fine in production without any issues.
Can anyone give a hint or suggestion what could be the issue.
Cheers,
Anjana Rao
11-01-2019 1:40 PM
With Oracle systems, I usually look at the shared cursor cache in ST04, select all statements about DFKKOP, it shows even the ones which are currently running, refresh to make sure which DFKKOP request corresponds to yours (probably yours reads intensively)... If so, it's probably the culprit and you may analyze its execution plan (what index and what strategy it's using, etc., compare with the one in production). Based on your analysis, some solutions can be the calculation of statistics, creation of missing indexes... Note that you also have a button to reach the corresponding ABAP code.
The difference with the production system is also maybe due to the volume of data which is processed and a slowest disk, RAM or CPU. Here, no other solution than reducing the volume processed, or just wait.
11-01-2019 6:55 AM
Hi Anjana,
Rather than hit & trial, please use new debugger feature effectively. This blog post has exact use case. I hope this helps:
https://blogs.sap.com/2019/10/10/abap-debugger-scripting-my-first-real-use/
GK
11-01-2019 7:16 AM
11-01-2019 1:40 PM
With Oracle systems, I usually look at the shared cursor cache in ST04, select all statements about DFKKOP, it shows even the ones which are currently running, refresh to make sure which DFKKOP request corresponds to yours (probably yours reads intensively)... If so, it's probably the culprit and you may analyze its execution plan (what index and what strategy it's using, etc., compare with the one in production). Based on your analysis, some solutions can be the calculation of statistics, creation of missing indexes... Note that you also have a button to reach the corresponding ABAP code.
The difference with the production system is also maybe due to the volume of data which is processed and a slowest disk, RAM or CPU. Here, no other solution than reducing the volume processed, or just wait.
11-05-2019 9:50 AM
Thanks Sandra.. This helped me identify indexes being used and see time difference between 2 systems to fetch records.. Although i am still in process of exact solution to help me identify issue in pre-production.. this was very helpful
11-01-2019 1:50 PM
Could be due to poor index quality in Oracle database. To rebuild it it is RSANAORA, but I cannot remember the program to calculate the quality (you could find it in db02_old transaction)
11-04-2019 10:22 AM
Moderator : I want to accept Sandra rossi comment... this helped me a lot.. can not find accept under her response.. can you help
11-04-2019 6:45 PM
It requires that I first convert my comment into an answer, then you can accept it. Let me do that.
If you have found a solution (or not), a feedback about the solution is welcome. Note that you may also post your own answer and accept it.
PS: Please use the COMMENT button for comments, questions, adding details, etc., ANSWER is only to propose a solution, dixit SAP text at the right of the answer area.