TSV_TNEW_PAGE_ALLOC_FAILED Error in Mass Update of Alternative Hierarchies

[Former Member]

Hi Experts,

I am updating Cost Centers in the alternative nodes by calling a standard Program

PERFORM hierarchy_generate IN PROGRAM saplkkhi.

All these Entries are coming from a file having a Cost Center and respective hierarchies

under which it should be updated.

Suppose in a file a Cost Center was attached to one node and if it needds to be detached

in to another node within same common node i am using a Function Module K_HIERARCHY_UPDATE

to update the cost center under given node.For all these Process i have created a Z Function

Module.

A report is executed in background which calls the above FM and was supposed to display all

the Cost Centers and Hierarchies attached.

Unfortunately I am getting TSV_TNEW_PAGE_ALLOC_FAILED Error even when i am Processing the

Report in Background.

Please help me solving this Issue.

Regards

Manjari Patankar

[nirajgadre]

Hi,

It seems it is related to some memory issue. Check with your basis team.

[former_member194613]

If there is a memory shortage, then you must take care to use less memory

As detailing are not given, I would assume that there are not enough COMMITs and all UPDATEs are kept to long in memory.

[yuri_ziryukin]

As an addition to the answer from Siegfried.

Very often SAP application developers define internal buffering tables (in ABAP) to store some information for the fast access. Initialization of these tables are normally programmed in some function modules that are called in certain scenarios. However I have seen the cases where initialization was not intended at all. Developers motivated it with the fact that corresponding functions are not designed to call them in batch many times or with large data volumes.

If such kind of initialization does not happen, you are going to have memory issues.

I would proceed as follows:

1. Run your job in foreground

2. In SM50/SM66 monitor the memory consumption of the work process.

3. When it goes high, debug the work process from SM50/SM66.

4. Being in debugger, look at the memory areas and find out objects that occupy a lot of memory.

5. Display these objects (normally internal tables) in ABAP editor and use "where used list" to find the places where they are initialized.

6. Try to call initialization functions/subroutines in your coding after each data package and subsequent commit work and wait.

[former_member192616]

As an addition to the answer of Yuri.

you may want to read note: 649327

Kind regards,

Hermann