SUMG related issue

[Former Member]

We are performing Unicode migration of ECC 6.0 EHP4 (701) system. We have migrated from ECC 6.0 EHP4 non Unicode to ECC 6.0 EHP4 Unicode. However we have issues in SUMG processing in target system.

1) In source system, spumg transaction code we have performed scans and maintained vocabulary. After completion of Reprocess log scan , we got around 98000 conversion collision entries in Reprocess log for 663 tables.

2) We selected affected row and displayed Reprocess log.

3) Its showing entry without any issues as shown below. If we logon to ZH language its showing as Chinese characters without any issueu2019s. As per Unicode conversion guide downloaded, we understand that these reprocess log entries can be fixed after Unicode migration in target system. So we have ignored these entries and exported database and imported into target system.

4) However, After we have built work list as per Unicode conversion guides and started sumg transaction code in target system (Unicode). We checked work list entries and tried to apply hints. But none of the existing or newly created hints working for work list entries. Then we have analyzed and tried to display entryu2026

5) Entry is being displayed as junk characters. Even we tried to display by logging in using other languages but its showing as junk characters. Its shows the entry as junk because of default fallback code page 1100-English has been assigned to all reprocess log entry in conversion (export / import). So our language experts are not able to identify these words for fixing in sumg in target system (Unicode).

Now we have below questions:

1) How do we proceed in this situation to fix sumg worklist entries?

2) Are these entries were supposed to be fixed in source system itself before we take export? If yes, after fixing how can we transfer these reprocess log entries for using in other system migration in landscape?

3)If these entries can be fixed in target system, how can we do it?

You help will me much appreciated..

[nils_buerckel]

Hi,

1) I strongly recommend to fix this problem in the source system. SUMG is designed for a limited number of entries only.

2) I would recommend to first try to resolve the collisions (and to fill the empty words) in the vocab which caused the reprocess entries. Then please maintain the reprocess logs as much as possible (hence you can use automatic repair in SUMG). If your SP is uptodate, you can even use hints in the reprocess logs via UMG_EXECUTE_HINT_FOR_REPLOG.

Please note: Maintaining reprocess logs is done during preparation phase (Uptime), SUMG has to be finished during downtime !

Transport of reprocess logs:

Please check the Unicode conversion guide. There you will find:

Begin(Excerpt)

Reuse Reprocess Logs

You can upload/download the language assignments of a Reprocess Log on a local PC or on a application server for later usage.

Go to Extras > Up/Download Reprocess Logs. Follow the instructions in the system documentation (select info icon from the toolbar).

End (Excerpt)

3) I would not recommend to do that (as stated under 1)

In general it is a big advantage if you know someone who has already done an MDMP conversion (e.g. make use of consulting).

Best regards,

Nils Buerckel