Some thoughts about services

I have a fresh new SAP ERP installation and one of the common tasks is automatically created jobs EU_INIT, EU_PUT, EU_REORG.

Here is my real-live shot.


1 500 000 seconds is equal to 416 hours!

Let’s look at the job log:


I hope someday the job will be finally completed.

But what am I talking about?

In my experience there is a widespread practice to use small frequent jobs. So sometimes I have to go to SM37 transaction and make sure that all is fine. For example, I would have 3-7 jobs recurring every five minutes (send mail, checks, post-processing tasks, receive data from external system, post data to external system and so on).

In this case daily view of SM37 is flooded with thousands of jobs that do nothing. BTW: Write logs to SLG1 is more practical solution than write to job log or spool.

Recurring short job can be converted into one big job with infinite loop inside:
do some work + wait five minutes + work again + sleep + work + … etc.

And here we come to general idea of services or daemons. I would like to implement this concept sometimes somewhere.

Error handling when calling other system using RFC


You have configured outbound connection and started to write code.

Iteration 1


Looks fine enough. But if you make mistake or connection is temporarily down or happens something else, you will get short dumps CALL_FUNCTION_REMOTE_ERROR,  CALL_FUNCTION_NO_DEST or CALL_FUNCTION_OPEN_ERROR etc.

You should always avoid short dumps in common environment.


Пасьянсы и установка SAP ERP

Установка SAP ERP – процесс местами странный и трудоёмкий.


Установка требует отдельной и жирной машины. Даже если это будет тестовая машина с минимумом данных для пары пользователей.

Для текущей стандартной версии мне понадобилась VM 12GB RAM/300GB HDD/вменяемый CPU.  И очень желательно, чтобы там ничего кроме сабжа не стояло и не планировалось.  Платформа SAP HANA имеет более высокие требования, непомерные для меня, хотя посмотреть бы хотелось. (далее…)