Error in Remote-Communikation | caught TrexNet exception

[jgleichmann]

Hello BIA experts,

we got this error 4 times and don´t know what´s the reason(from RSPC):

12.01.2009 15:33:57 Loading data to index BWP_BI0:SMAT_PLANT (records '0000318655', job '0' )

12.01.2009 15:44:16 Characteristic 0MAT_PLANT: 0 data records in version M deleted from table /BI0/QMAT_PLANT

12.01.2009 15:44:16 Index for table '/BI0/SMAT_PLANT' is being processed

12.01.2009 15:44:16 Index '/BI0/SMAT_PLANT' for BIA index deleted

12.01.2009 15:44:16 Index 'BWP_BI0:SMAT_PLANT' for BIA index created

12.01.2009 15:44:16 Execute rollback

12.01.2009 15:44:16 Rollback for index 'BWP_BI0:SMAT_PLANT' executed

12.01.2009 15:44:16 Runtime (ms): RFC server:58, BIA client:58, BIA Kernel: 0, ABAP RFC: 60

12.01.2009 15:44:16 Error in Remote-Communikation with partner http://[blade02]:30003/indexCellTable

12.01.2009 15:44:16 Job cancelled after system exception ERROR_MESSAGE

In the rfc alert trace there is no relevant error at this time. This error is independent from the blades because i moved the affected index to another blade. The result was the same.

I only can see an error in the IndexServerAlert trace:

2009-01-12 15:34:42.475 e SERVER_TRACE TRexApiIndexDocuments.cpp(09363) : start rollback bwp_bi0:smat_plant user=[user]

2009-01-12 15:34:42.489 e Trex_SE FuzzyOptimizer.cpp(04135) : start rollbackOptimize for index bwp_bi0:smat_planten

2009-01-12 15:34:42.508 e Trex_SE FuzzyOptimizer.cpp(04229) : rollbackOptimize for index bwp_bi0:smat_planten finished rc=0

2009-01-12 15:34:42.518 e SERVER_TRACE TRexApiIndexDocuments.cpp(09397) : finished rollback bwp_bi0:smat_plant rc=0

2009-01-12 15:34:53.213 e Service TrexService.cpp(01283) : caught TrexNet exception in method : Protocol//request begin identifier is invalid

In the syslog on BI I can see this error:

15:44:14 BWP82 BTC 007 010 BWBATCH TR 1 TREX_ICM-Receive_Error; return code = 402, return text = ICM_HTTP_TIMEOUT

I don´t know why I get here an ICM error, the BIA is so configured that it just uses the RFC connection to the backend system. There is also no error in the ICM traces in SMICM.

I only found a note(1092151) but this note doesn´t match to us because we already have SP15 in the backend system.

Thanks in advance!

Best Regards,

Jens

Edited by: Jens Gleichmann on Jan 13, 2009 8:15 AM

Edited by: Jens Gleichmann on Jan 13, 2009 11:52 AM

[former_member93896 · Accepted solution]

Hi Jens,

ICM is used to index SID tables. Your ICM in probably not configured properly or can not access all blades. Please see SAP note <a href="http://service.sap.com/sap/support/notes/1102652">1102652</a>.

Regards,

Marc

SAP NetWeaver RIG

[jgleichmann · Accepted solution]

Hi Marc,

thanks for your reply.

Yes, I already found this note, but I can´t found a problem in the configuration of the ICM. It´s very strange if we rerun the process chain than it works. It´s not only one index which has the error, there are now 3 indizes which have the error.

This process chain also run in the year 2008 and from my team didn´t change anything on BIA or ICM.

In SMICM I can´t find any log about this errors. Is this actually logged by the ICM? Does it help if I set the trace level higher?

The error can be reproduced at all 4 blades.

Here an other error:

Prozess 'Initiales Aktivieren und Füllen von BIA-Indizes ',

Variante 'BIA-Index aktivieren und initial füllen für MM_I_I' hat

Stat us 'fehlerhaft beendet ' (Instanz

'BIBIA4CI394JSTM6Q9TLJG99P5PNHC ')

in English:

process 'initial activate and fill of bia-indizes'

variant 'BIA-index activate and fill for MM_I_I' have

status 'finished with error' (Instance

'BIBIA4CI394JSTM6Q9TLJG99P5PNHC ')

Best Regards,

Jens

Edited by: Jens Gleichmann on Jan 15, 2009 7:47 AM

[jgleichmann · Accepted solution]

I have solved the problem. For anyone who has the same problem:

- increase the http timeout value ( in my case I set it to 1800 )

- recreate the index by following the note 1098260

That´s it!

Best Regards,

Jens

[Former Member · Accepted solution]

Jens,

How did you increase http timeout value?

which parameter , OSS note?

Thanks,

Saurabh Bansal

[jgleichmann · Accepted solution]

Hi Saurabh,

goto transaction TREXADMIN then click on TREX-ABAP-Customizing(not F8 for the actually TREXADMIN). Then you can see a parameter called "HTTP-Timeout ", set it to 1800. I didn´t find a recommended value in a note, but I asked SAP and they told me that it could be set to 1800.

Best Reagrds,

Jens

[Former Member · Accepted solution]

Jens,

Is it the same error that you discussed in this thread.

P.S. - Ours is already set to 1800. thanks for the info though.

-Saurabh Bansal

[jgleichmann · Accepted solution]

Hi Saurabh,

no, it´s not the same error, because the kernel was updated since this problem occured and this solved it.

Best Regards,

Jens

[Former Member]

We had such a problem. We decided to reboot the server.