Error in phase RUN_RSPTBFIL_PDIST while upgrading from BW 3.5 to BI 7.3

[Former Member]

Hello

I'm having a very serious problem when I try to upgrade from

BW 3.5 to BI 7.3 SP04.

There's no documentation about how to solve it nor in sdn neither in service.sap.com.

I did the prerequisite steps as stated in the upgrade guide and

associated notes but I'm having the following error in the module phase

MAIN_SHDRUN/RUN_RSPTBFIL_PDIST.

The log file RSPTBDST.ELG shown this error

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

PUTTB_SHD post distribution actions ERRORS and RETURN CODE in RSPTBPOD.WL5

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

2EETG011 "[DIST_CHK] conversion not possible for CLU4"

2EETG011 "[DIST_CHK] conversion not possible for VER_CLUSTR"

The program executed in this PHASE is : RSPTBFIL

The form where the problem appears is : dist_chk

The error shows up when the lines marked below like this "--HERE IS TE ERROR>>" are executed

-

---

PERFORM write_log_line USING ' ' '011' '[DIST_CHK]'

'Checking for conversions for physical clusters.' '' ''. "#EC NOTEXT

"--HERE IS TE ERROR>>" SELECT * FROM ddxtt AS x INTO TABLE lt_ddxtt

WHERE modeflag = 'U' AND

tabtype = 'C' AND tabform = 'T'

AND EXISTS ( SELECT * FROM tbatg WHERE object = 'SQLT' AND

tabname = x~tabname AND fct = 'CNV' ).

DESCRIBE TABLE lt_ddxtt LINES ld_cnt.

IF ld_cnt > 0.

LOOP AT lt_ddxtt INTO ls_ddxtt.

ld_logline = lv_line_message.

"--HERE IS TE ERROR>>" REPLACE '&1' WITH 'conversion not possible for' INTO ld_logline.

"#EC NOTEXT

REPLACE '&2' WITH ls_ddxtt-tabname INTO ld_logline.

CONDENSE ld_logline.

PERFORM write_log_line USING 'E' '011' ld_logline '' '' ''.

ENDLOOP.

ELSE.

PERFORM write_log_line USING ' ' '011'

'[DIST_CHK]' 'No physical cluster conversion found' '' ''. "#EC NOTEXT

ENDIF.

ENDFORM. " dist_chk

-

---

When i executed transaction SE14-> DB Requests->created by import

I saw (amongst others) this two entries:

Object Object name ID Function Order XID User Date Error

SQLT CLU4 CNV 0 I DDIC 14.09.2011

SQLT VER_CLUSTR CNV 0 I DDIC 14.09.2011

My source version is

SPAM version: 640/0043

SAP_ABA 640 0019 SAPKA64019

SAP_BASIS 640 0019 SAPKB64019

PI_BASIS 2005_1_640 0009 SAPKIPYJ69 Basis Plug-In (PI_BASIS) 2005_1_640

ST-PI 2008_1_640 0001 SAPKITLRC1 SAP Solution Tools Plug-In

SAP_BW 350 0019 SAPKW35019

BI_CONT 353 0007 SAPKIBIFP7

ST-A/PI 01L_BCO640 0000 - Servicetools for other App./Netweaver 04

The SAPup Version I'm using is the last SAP has released:

The current SAPup release and version are:

SAPup release lmt_002, version 41.000

Does anyone know how to fix or avoid this issue?

Kind regards

[former_member189725]

Refer SAP note Note 902817 - Inconsistencies in cluster tables (log vs phys field name)

in such a scenario.

Regards

Ratnajit

[Former Member]

Hello All,

I found the solution for my particular situation!

When checking the runtime object for CLU4, an inconsistency in the runtime object popped up:

- the ABAP dictionary contains the field K5N1

- the DB dictionry contains the field name K5N01

This is a known and documented inconsistency, see note 902817 for the description.

BUT: the solution pointed out there is wrong! The note suggests to delete all content from the physical tables and then activate the table in SE14. This is not possible for cluster tables neither in SE14 nor in SE11.

This solution worked for me for both CLU4 and VER_CLUSTR:

- as the tables should be empty normally I dropped them in the database

- unlock the shadow instance with SAPup

- log on on to the shadow instance

- I then made some arbitrary change to the table definition in SE 11 (like adding a field). Make sure to maintain the size class also, otherwise the table won't be activated

- save the changes and activate the tables

- take back the previous change

- save and activate again

- lock the shadow instance

- restart the upgrade phase

Once the upgrade is done, check whether the physical table exists in the DB.

Hope this helps!

Regards,

Matthias.

Edited by: Matthias Dörfel on Nov 18, 2011 3:14 PM

[Former Member]

Hi Matthias

Thanks for sharing the solution.

Best Regards

[Former Member]

Hallo Matthias,

what do you mean by: "Make sure to maintain the size class also, otherwise the table won't be activated".

Where can I maintain the size class?

I tried to add a field to the cluster tables CLU4 and VER_CLUSTR, but was not able to activate them.

Thank's in Advance,

Daniel

[Former Member]

In case it helps someone:

I had the same problem during an upgrade from NW702 to NW73.

I didn't read the note NW73 upgrade note carefully enough and during ASU-Toolbox, step clean inconsistencies (db02) i created the missing database tables in the source-system.

during the upgrade with sapup i received the mentioned inconsistency errors about CLU4 & VER_CLUSTR

Also I got errors when activating the tables as mentiond by matthias.

The solution was like Matthias said, but a little different (bold):

- unlock the shadow instance with SAPup

- log on on to the shadow instance

- create a copy of DDIC user

- login as normal user

- change system change mode to modifiable (SE06)

- in se11, change the name of the first field to:

CLU4: K1N05

VER_CLUSTR: MANDT

- (save and) activate

- lock the shadow instance

hope it helps

[Former Member]

Hi All,

I am experiencing a very similar problem during upgrade from NW04 / 6.40 to 7.01: SAPup fails in phase MAIN_SHDRUN/RUN_RSPTBFIL_PDIST with an error BATCH JOB RSPTBFIL FAILED. THis is the output of RSPTBDST.ELG:

-

---

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

PUTTB_SHD post distribution actions ERRORS and RETURN CODE in RSPTBPOD.T1B

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

2EETG011 "[DIST_CHK] conversion not possible for CLU4"

2EETG011 "[DIST_CHK] conversion not possible for VER_CLUSTR"

-

---

The log file RSPTBPOD.<SID> details the error:

-

---

4 ETG011 "-- POSTDIST DIST_CHK RSPTBPOD.<SID> --"

4 ETG011 "[DIST_CHK]""Checking for conversions for physical clusters."

2EETG011 "[DIST_CHK] conversion not possible for CLU4"

2EETG011 "[DIST_CHK] conversion not possible for VER_CLUSTR"

4 ETG011 "-- POSTDIST DIST_MOD RSPTBPOD.<SID> --"

-

---

I can find identical records in SE14 --> DB Requests ... however, these two tables showd up earlier in SPDD and we adjusted them to go back to SAP standard.

Anyone any updates on this issue?

Thanks and regards!

[Former Member]

Hello,

I have seen these kind of errors during unicode conversion. And I think, before upgrade too, we should check cluster tables for incorrect records. In that case, you should have gone through the Note 1077403 - Cluster table check using SDBI_CLUSTER_CHECK before starting upgrade.

Please go through some similar notes, may be some hint for your problem ?

Note 1052298 - Inst.NW 7.0 SR3/Business Suite 2005 SR3-UNIX

Note 1341276 - Inst.NetWeaver 7.0 EHP2/Business Suite 7i2010-UNIX

Thanks

[Former Member]

Thank you very much for the info but those notes does not apply in my case.

Both notes deals with unicode problems, that's not my case.

I'm just trying to upgrade.

Thanks anyway.

[nils_buerckel]

Hi,

did you check whether the named tables exist on the DB ?

E.g. via SE11 > Utilities > DB objects > Check ?

Best regards,

Nils Buerckel