R3check for cluster table contains incorrect cluster record

[Former Member]

Hi all,

We are doing unicode conversion on our current Solman 7.0 SP16 OS NT system.

We run UMG_R3CHECK_FILES to check cluster tables we referred to Note 89384 -createt control for this cluster table that contains incorrect cluster records.

But when we on os level run command R3check CLU4_200 > OUT_CLU4_200 the control file

have not this information on output file <phys. cluster table>: <key1><key2><key3>... (<error number>) referred to Note 89384

But we get this message on output file:

This is R3CHECK version 6.14 (release 700 - 17.02.10 - 13:07:00).

SAP myCluster Technology (dbcheckclu - Mar 15 2010 - 20:00:54).

Current time is 11.01.2011 08:38:06,40h GMT.

R3CHECK finished (0000).

Can you please guide us the steps for cluster table check with R3check

Regards

[Former Member]

Same problem here, Running Solaris 10, IBM DB2 9.5 LUW FP4, R3 4.6C.

SDBI_CLUSTER_CHECK program reports Table RFBLG has 0 initial record but with 40 other errors.

The good thing is that the errors shown, reference records for a client that is not a productive client in the system.

The unicode conversion was completed.

Raised an OSS in the respective component and awaiting response.

Btw, is it possible to find out from the system, who are the table members for cluster table RFBLG?

[Former Member]

Problem not solvet but i clos it. The problem is with Oracle DB.

thanks for help.

[tomas_black]

Hello Ejersbo,

the message you got actually informs you that there are no errors on the output file, so the table(s) is OK from the cluster interface point of view.

The message

"This is R3CHECK version 6.14 (release 700 - 17.02.10 - 13:07:00).

SAP myCluster Technology (dbcheckclu - Mar 15 2010 - 20:00:54)."

tell us that the cluster interface was entered, thus the R3check control file was ok.

This is good news actually, you've got no errors

Best regards,

Tomas Blacsk

[Former Member]

Hi Tomas,

Thanks for reply. We have a solman 7.0 sp16 that it is not unicode on OS 32bit.

I want to unicode this installationand and export it on a OS 64bit installation.

When I run TC- SPUMG and check Consistency it look lik it is working.

But when I chech under Extras -> Cluster information find A list of table cluster with initial records.

Så it is strang if I have not any errors after R3check. When I try by SAPINST creat a system copy I get error with this cluster table error.

Is it possible to delete this cluster tables?

Regards

[former_member227281]

Hi,

Is it possible to dfelete cluster tables: The Answer is Yes. It is possibler. However the SAP system will be completely crashed if you delete cluster tables and SAP will not work properly. All data will be inconsistencies.

Hence SAP doers not recommend to delete cluster tables.

Please send us the error message when you execute SAPINST command to create System copy .

Thanks and Regards

APRao

[Former Member]

Hi A Prasad Rao,

Thanks for replay.

Her is my error from SAPINST os system copy.

On Export ABAP I get 1 failed. The log is:

From export_monitor.log file:

TRACE: 2011-01-18 12:29:30 com.sap.inst.migmon.LoadTask processPackage

Unloading of 'DOKCLU' export package from database:

slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe -e DOKCLU.cmd -datacodepage 4103 -l DOKCLU.log -stop_on_error

TRACE: 2011-01-18 12:29:30 com.sap.inst.migmon.LoadTask processPackage

Unloading of 'ALALERTDB' export package from database:

slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe -e ALALERTDB.cmd -datacodepage 4103 -l ALALERTDB.log -stop_on_error

.

TRACE: 2011-01-18 12:30:00 com.sap.inst.migmon.LoadTask processPackage

Unloading of 'DSVASREPODOCS' export package from database:

slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe -e DSVASREPODOCS.cmd -datacodepage 4103 -l DSVASREPODOCS.log -stop_on_error

ERROR: 2011-01-18 12:52:38 com.sap.inst.migmon.LoadTask run

Unloading of 'SAPCLUST' export package is interrupted with R3load error.

Process '
slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe -e SAPCLUST.cmd -datacodepage 4103 -l SAPCLUST.log -stop_on_error' exited with return code 2.

For mode details see 'SAPCLUST.log' file.

Standard error output:

sapparam: sapargv( argc, argv) has not been called.

sapparam(1c): No Profile used.

sapparam: SAPSYSTEMNAME neither in Profile nor in Commandline

For mode details see 'SAPCLUST.log' file.

Standard error output:

sapparam: sapargv( argc, argv) has not been called.

sapparam(1c): No Profile used.

sapparam: SAPSYSTEMNAME neither in Profile nor in Commandline

---

Her is a port of my SAPCLUST.log file:

(RTF) ########## WARNING ###########

Without ORDER BY PRIMARY KEY the exported data may be unusable for some databases

slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe: job completed

slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe: END OF LOG: 20110118125230

slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe: START OF LOG: 20110118125231

slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe: sccsid @(#) $Id: //bas/700_REL/src/R3ld/R3load/R3ldmain.c#21 $ SAP

slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe: version R7.00/V1.4

Compiled May 3 2010 23:30:54

slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe -e SAPCLUST.cmd -datacodepage 4103 -l SAPCLUST.log -stop_on_error

(DB) INFO: connected to DB

(DB) INFO: DbSlControl(DBSL_CMD_NLS_CHARACTERSET_GET): WE8DEC

(EXP) INFO: check NameTab widths: Ok.

(DB) INFO: Export without hintfile

(RSCP) INFO: UMGCOMCHAR read check, OK.

(RSCP) INFO: environment variable "I18N_POOL_WIDTH" is not set. Checks are active.

(RSCP) INFO: I18N_NAMETAB_TIMESTAMPS not in env: checks are ON (Note 738858)

(RSCP) WARN: UMGCONTAINER has 1 more problems.

(RSCP) WARN: UMGSETTINGS nametab creation: status unknown.

(RSCP) INFO: Global fallback code page = 1160

(RSCP) INFO: Common character set is not 7-bit-ASCII

(RSCP) INFO: Collision resolution method is 'fine'

(RSCP) INFO: R3trans code pages = Normal

(RSCP) INFO: EXPORT TO ... code pages = Normal

(RSCP) INFO: Check for invalid language keys: activated

(RSCP) INFO: I18N_NAMETAB_NORM_ALLOW = 999999999

(RSCP) INFO: I18N_NAMETAB_NORM_LOG = 1000000002

(RSCP) INFO: I18N_NAMETAB_ALT_ALLOW = 10000

(RSCP) INFO: I18N_NAMETAB_ALT_LOG = 10003

(RSCP) INFO: I18N_NAMETAB_OLD_ALLOW = 0

(RSCP) INFO: I18N_NAMETAB_OLD_LOG = 500

(RSCP) INFO: init SUMG interface (3#$Revision: #22 $); package name: "SAPCLUST".

(RSCP) INFO: "SAPCLUST001.xml" created.

(GSI) INFO: dbname = "SLM20060713110741"

(GSI) INFO: vname = "ORACLE "

(GSI) INFO: hostname = "SLMNT1 "

(GSI) INFO: sysname = "Windows NT"

(GSI) INFO: nodename = "SLMNT1"

(GSI) INFO: release = "5.2"

(GSI) INFO: version = "3790 Service Pack 1"

(GSI) INFO: machine = "2x Intel 80686 (Mod 15 Step 1)"

(GSI) INFO: instno = "0020210560"

-

(RTF) ########## WARNING ###########

Without ORDER BY PRIMARY KEY the exported data may be unusable for some databases

(EXP) INFO: table CDCLS will be exported with sorting

myCluster (63.25.Exp): 5575: damaged 0. cluster record (pgs_total 0, bitem_total 0)

myCluster (63.25.Exp): 3818: received return code 2 from decompression routine.

--

myCluster (63.25.Exp): 2887: failed to unpack cluster item.

*myCluster: CDCLS *200***SADOCATTR ***DDEE0011154442F1A77C000C292175B7 ***0000024642*

myCluster (63.25.Exp): 329: error during conversion of cluster item.

myCluster (63.25.Exp): 332: affected physical table is CDCLS.

(CNV) ERROR: data conversion failed. rc = 2

(DB) INFO: disconnected from DB

--

slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe: job finished with 1 error(s)

slmnt1.dlg.dk\sapmnt\SLM\SYS\exe\run\R3load.exe: END OF LOG: 20110118125238

Any help

regards

[tomas_black]

Hello Ejersbo,

first: like Prasad Rao said, do not ever delete a cluster table in a SAP system unless explicitly stated by SAP Support (or SAP personnel). It's quite complex to technically explain why but in the end you'll have to perform a full restore from your last valid backup in order to restore the system to a consisten state again.

Now regarding this error seen in sapinst: it's actually quite a tricky one. It says the affected table is CDCLS; is there a file called CDCLS.log available? Please post the content of this file on this thread. If not, is there a file called RFBLG.log? Do the same.

The error occurs just after the R3load command + codepage set up for some tables, which lead us to believe the problem is within the codepage used. As per SAP note #1066404 ("R3load, ORDER BY, Unicode conversion of cluster tables"), this error is not supposed to happen during import. The note states that the export must be repeated and the correct code pages must be

transferred to the R3Load.

Did you make sure that the correct code pages were chosen?

Also, please refer to note #765475 ("Unicode Conversion: Troubleshooting) -> UCTroubleshootingguide.zip

I hope this information helps fixing the problem.

Best regards,

Tomas Black

[Former Member]

Hi Thomas,

Thanks again for fast reply.

I am not delete any cluster table in sap. And there is no CDCLS.log or RFBLG.log file.

I did create a control file for CDCLS table by usning SDBI_CLUSTER_CHECK report:

when I run SDBI_CLUSTER_CHECK_ERRREPORT report

I get this messages.

SDBI_CLUSTER_CHECK: Report errors to SAP (note 1077403)

Table CDCLS : Number of erroneous records: 1 ( 0 initial, 1 other errors)

Key: MANDANT=200,OBJECTCLAS=SADOCATTR,OBJECTID=DDEE0011154442F1A77C000C292175B7,CHANGENR=0000024642

Errortext: ERR_READ - Return code from db_setget.(3 )

Timestmp: LO49ef((AAJ(D3DE0D6EAF558686F1918F000C292175B7

-

---

When I on OS command run R3check for CDCLS cluster table the contro file for this table is CDCLS_200.

I get this massage on out put file.:

This is r3check version 6.14 (release 700 - 17.02.10 - 13:07:00).

SAP myCluster Technology (dbcheckclu - Mar 15 2010 - 20:00:54).

Current time is 18.01.2011 11:59:16,68h GMT.

CDCLS : 200SADOCATTR DDEE0011154442F1A77C000C292175B7 0000024642(E000)

r3check finished (0000).

Regards

[tomas_black]

Hello Ejersbo,

-

---

EDITED in 19.01.2011: in fact, Development needs to check the CDCLS records, he believes the record is corrupted. You should open an OSS message in SAP Support reporting this problem, under component BC-DB-DBI. You can refer to this thread for the initial steps already taken. I was taking a look at similar msgs and most of them end up in Dev Support with detailed analysis (and solutions) for this type of problem... My apologizes for not being able to help you fixing this issue.

-

---

I'll edit this reply message afterwards - I'm contacting the owner of note #1077403 to properly advise the best course of action for this issue.

For now I can only tell you that the error ERR_READ means that a cluster record couldn't be read, but I'm not sure what's the root cause yet. In the meantime, please analyze how many entries in CDCLS are corrupted by using the report RSSCD_CHECK_CDHDR to create a list of headers without positions. This report might take a long rtime, depending on the

number of entries in table CDHDR, so run it in backgroud. You can also run report RSSCD_CHECK_CDPOS - please refer to note #1023737 for more information. If a lot of entries are found you can also refer to notes #1236778 and #1245574.

I'll edit this message later on with further information, when I get the Developer advise.

Best regards,

Tomas Black

Edited by: Tomas Black on Jan 19, 2011 2:25 PM