Urgent: Unicode conversion - table splitting

[ThomasKjelsrud]

Hi all,

I am having a problem when trying to perform the export step of the unicode conversion on ERP2005, MSSQL server 2005. Due to previously very long runtime, I am trying to use the table splitting option. I have performed the "Table Splitting Preparation" step with what seems to be a success.

The problem is when I run the actual database instance export.

First of all:

Which filepath should I provide SAPinst when it is asking for "Table input file"?

Second:

How can I actually determine the Package Unload Order? I tried selecting this option, but I was not given the opportunity to change this order in the subsequent screen. (The F1 help in SAPinst sais that I would be able to...)

Have anyone of you experience with this?

Best Regards,

Thomas

[Former Member]

Hi Thomas,

As I can imagine from the date of your last posting, you probably have your answers. But just so other people, who are searching on this topic, can find the answer here, I will fill in the blanks.

First, when using table splitting, you must export using the migration monitor (MIGMON). That means, when you run SAPINST, on the ABAP System -> Database Export screen, select the Export Method: Export using Migration Monitor.

At this point, as you said, you have already run the Table Splitting option from SAPINST, so the WHR files are in the export DATA directory. This whr.txt file is also there.

The file input screen you are mentioning does not appear when you select the Export method mentioned above. But you can create such a file for MIGMON. You have to create this on your own and can only do it after the SAPINST has split the STR files.

When finished with SAPINST, create the .txt file (I usually call it table_order.txt) and add the names of the STR files which you want to make sure are exported first. After MIGMON completes that list, all other STR files are exported in alphabetical order.

Create your table_order.txt file inserting the filenames of the packages that were created, without the STR/WHR extensions (using some large tables which STR splitter broke out into their own STR files during my last export). You have to look in DB02 and sort descending based on used space to determine the order in which these tables should be exported and listed in the .txt file:

<SPLIT Table>-1

<SPLIT Table>-2

.

.

.

<SPLIT Table>-n

BSIS

RFBLG

CE1VVOC

CE1WWOC

DBTABLOG

CE3VVOC

COEP

ARFCSDATA

GLPCA

KONV

SWW_CONT

SOC3

CDCLS

CE3WWOC

BSE_CLR

STXL

EDIDS

COSP

BSAD

EDI40

ACCTIT

BSIM

VBFS

BSAS

ACCTCR

CDHDR

CE4VVOC_ACCT

SGOHIST

MSEG

In the export_monitor_cmd.properties file, you specify this file name for the 'orderBy=' parameter. Create and keep the table_order.txt file in the same directory as the export_monitor.sh/.bat files.

For importing, again you will need to use the Migration Monitor. SAPINST will automatically stop and prompt you to start the import using MIGMON. Here, you can specify the same table_order.txt file. You might want to amend it to control when the rest of the packages are imported, if you found one or more tables holding up the completion of the export.

I hope this helps someone.

Best Regards,

Warren Chirhart

[ThomasKjelsrud]

Getting help from SAP AG

[Former Member]

Hi Thomas,

We recently did a unidode conversion of our BI 7 system running on Oracle 10 G.At the first trial on test server we have used the splitting option but we got stuck when we did the import [We did not change the package unload order we user the defult one].

So what happened is though import went successful at the point of starting system we got stuck due missing entry in TADIR.

After searching notes we came to know the issue was the sptitting of package containg DDNTT and DDNTF tables.

Finally we completed the process by redoing the export without splitting the package and it went successfully [Our DB size was 373 GB during export]

The note 952514 - Using the table splitting feature

What is your DB size.Based on my exp what i feel is if your db size is less than 500 GB you don't have to split the packages.

[ThomasKjelsrud]

Hi,

Our database size is approx 700 GB for Test and about 800 GB for Production, so the runtime is unacceptable for us.

We are trying to split the SOFFCONT1 table, I do not believe that there are any DDNTT or DDNTF tables involved here.

I created a "split.txt" file, containing SOFFCONT1%2, and supplied this file during the "Table Splitting Preparation" in SAPinst. This process gave me three files as output: wht.txt, SOFFCONT1-1.whr and SOFFCONT1-2.whr. When asked to provide the Table Input file during Database instance export in SAPinst, I provided the wht.txt file. Do you know if this is the correct file?

Thanks,

Thomas

Are there really no one that have any experience with table splitting during unicode conversion??

Message was edited by:

Thomas Kjelsrud