Very often, when working in SAP support, we receive incidents of users asking us to find out the root-cause of an issue that happened in the past but that is not happening anymore. The answer for this kind of inquiry is usually "please provide a reproducible example of the issue on your system" and frequently the person who opened the incident cannot provide this example, but insists to know the root cause of the issue.
Let me explain what is our troubleshooting line on an application issue, in order to show how important is this reproducible example. When an incident is received by SAP Support, the following steps are carried out by the person assigned to solve the incident:
1 - General research on the documentation: In order to understand if the issue described is according to the system standard design or if there is a customizing to change the system behavior, the first step is to research on the documentation for a better understanding of the scenario. This research may include SAP Help, wikis, SCN or internal documentation and this step may be skipped if the scenario is already known.
2 - Notes search: If the research on the documentation points that the behavior described by the user is not planned, the second steps is to search for notes that may fix the issue or explain the scenario from a technical point of view. This notes search includes notes with code correction, modification notes, consulting notes, FAQ notes and KBAs.
3 - Search for similar problems: If there are no reference on the documentation or on notes, the next steps is to make a research in an internal database for similar issues described in the past. This is a very useful source of information, and we can find details about the analysis already made on a similar issue by another support engineers or developers.
If none of the above mentioned alternatives bring a solution, then the next steps is to analyze the issue directly on the system where it is happening. In this case, the best way to do it is to reproduce the scenario in the SAP internal test systems and then compare with the scenario on the system where the issue is happening.
When there is no master data or customizing setting difference that could explain the cause of the issue, there is only one option
we must run a detailed investigation using the debugger, to analyze the program execution and to understand why there is a difference between the SAP internal test systems and the system where the issue happens. In most cases, it is necessary to debug both systems in parallel, in order to understand where is such difference.
Below there are some very common issues, that it is generally only possible to find out the cause of the issue analyzing a reproducible example in debug:
1 - Database inconsistencies: Generally, a database inconsistency is only identified after it already happened. In this case, SAP can generally provide a correction report to fix the inconsistency. Analyzing the inconsistent data on the database tables is not enough to find the root cause of the issue, because the same table can be update by many different programs or transactions. Running the transaction without being able to recreate the inconsistency it is also not enough, since it will be only possible to observe the standard program flow.
2 - Issues caused by user-exits, BAdIs, modifications or customer enhancements: Very frequently, the issue is caused by a custom code implemented on an exit, BAdI, modification or enhancement. In this case, it is also necessary to analyze a reproducible example in debug, since the custom code implemented is not known to SAP. By running a reproducible example on the customer system, it is possible to jump the custom code in debug, in order to identify if the issue lies on the standard SAP code or on the custom code.
3 - A short dump or any issue that happens randomly: It is very common that an issue, such as a short dump, happens randomly on the system and the user cannot track exactly the steps to recreate this short dump. In some cases, the short dump provides useful information that helps to identify what is causing it, but this is not always true. Sometimes, the short dump does not provide enough information and it is necessary to reproduce the issue, so that it is possible to see the values of the internal variables and how they are calculated.
4 - A problem that happened in the past and it does not happen anymore: It is usual that an issue is observed on a system once and then it never happens again. Considering my area of expertise, I usually hear that the MRP results are incorrect, but when I try to run MRP again for the same material, everything is perfect. Unfortunately, it is not feasible to keep logs of everything on the system, as it would increase the database size and lead to performance issues. Therefore, sometimes it is not possible to interpret the results of an MRP run that happened in the past. The same is valid for all the areas.
If you take a look on the SAP 9, it explains that the when the issue is not reproducible, it may not be always possible to find out the root cause of the issue.
Therefore, before opening an OSS incident to SAP, please try by all the means to track the steps to reproduce the issue on your system. If this is not possible, please be comprehensive if you are asked for a reproducible example and try to understand if it is not possible to find out the root cause of the issue without this example.
