TREX Preprocessor Error

[peter_langner]

Hello,

we are using TREX 7.1. When I try to index documents it does not work. In the tracefile of the preprocessor I found the following:

"[4832] 2008-09-23 15:37:05.665 e preprocessor Preprocessor.cpp(01161) : HTTPHEAD failed for URL http://S0001010.srz0.loc:8010/sap/bc/contentserver/100?get&pVersion=0045&contRep=BDS_DB23&docId=DC8C... : Errorcode -2 , Message Reader::connect failed, url=http://S0001010.srz0.loc:8010/sap/bc/contentserver/100?get&pVersion=0045&contRep=BDS_DB23&docId=DC8C4FD0F51DD5F196B10015171E5D46&compId=Tchibo.pdf&accessMode=r&authId=CN=EDS&expiration=20080927133703&secKey=MIHzBgkqhkiG9w0BBwKggeUwgeICAQExCzAJBgUrDgMCGgUAMAsGCSqGSIb3DQEHATGBwjCBvwIBATATMA4xDDAKBgNVBAMTA0VEUwIBADAJBgUrDgMCGgUA

oF0wGAYJKoZIhvcNAQkDMQsGCSqGSIb3DQEHATAcBgkqhkiG9w0BCQUxDxcNMDgwOTIzMTMzNzAzWjAjBgkqhkiG9w0BCQQxFgQUDUoHRNCbcSTa3T9yhP/S6xtGqqcwCQYHKoZIzjgEAwQwMC4CFQCNBo3JCqFypSC0jcu7GtYb1GdgIVAKhj

[4832] 2008-09-23 15:37:05.665 e preprocessor Preprocessor.cpp(03965) : HANDLE: DISPATCH - Processing Document with key 'BDS_POC23 DC8C4FD0F51DD5F196B10015171E5D460000000000000000000000' failed, returning PREPROCESSOR_ACTIVITY_ERROR (Code 8002)

[5608] 2008-09-23 15:37:05.665 e preprocessor Preprocessor.cpp(01161) : HTTPHEAD failed for URL http://S0001010.srz0.loc:8010/sap/bc/contentserver/100?get&pVersion=0045&contRep=BDS_DB23&docId=DC8C... : Errorcode -2 , Message Reader::connect failed, url=http://S0001010.srz0.loc:8010/sap/bc/contentserver/100?get&pVersion=0045&contRep=BDS_DB23&docId=DC8C71E0F0A3B5F196B10015171E5D46&compId=edeka.txt&accessMode=r&authId=CN=EDS&expiration=20080927133703&secKey=MIHxBgkqhkiG9w0BBwKggeMwgeACAQExCzAJBgUrDgMCGgUAMAsGCSqGSIb3DQEHATGBwDCBvQIBATATMA4xDDAKBgNVBAMTA0VEUwIBADAJBgUrDgMCGgUA

oF0wGAYJKoZIhvcNAQkDMQsGCSqGSIb3DQEHATAcBgkqhkiG9w0BCQUxDxcNMDgwOTIzMTMzNzAzWjAjBgkqhkiG9w0BCQQxFgQUEdAW4SxVhKQpejFMlpENUExoZ4wCQYHKoZIzjgEAwQuMCwCFE1FligZ1Yu90EfxaxjODGshsQyAhQwmMKz7

[5608] 2008-09-23 15:37:05.665 e preprocessor Preprocessor.cpp(03965) : HANDLE: DISPATCH - Processing Document with key 'BDS_POC23 DC8C71E0F0A3B5F196B10015171E5D460000000000000000000000' failed, returning PREPROCESSOR_ACTIVITY_ERROR (Code 8002)" etc. etc.

If I drop the url into my browser, the document to be indexed is fetched properly. So what do all the errorcodes mean? What is the meaning of "HTTPHEAD failed for URL"?

Has anyone of you an idea?

Thank you for your help in advance,

Peter

[MarcelRabe]

https://www.sdn.sap.com/irj/scn/message?messageID=6233297

[peter_langner]

Hi Marcel,

this is the number of the actual thread. Please put the link to the new threat here and I will try to help you.

Peter

[MarcelRabe]

Strange....i updated the post with the correct link yesterday and now it's back again to the wrong one. Anyway here it is: https://forums.sdn.sap.com/message.jspa?messageID=6682123#6682123

[MarcelRabe]

you're right I'll create a new thread.

But the 6300 documents are not really indexed as they are links to help.sap.com html files and TREX has no proxy configured. But an index is created on trex, so is the queue. The files go through the queue and the preprocessor and have status indexed. But if I try to index a single document (one of the remaining 25.000) via transaction SKPR07 I get the message "Error Indexing not allowed when marking indexing for document SOLARGNSRC, 0010503D4800FA59E10000000A114AB6"

There is no explainantion why this error occurs or what it means. No log either.

Marcel

[peter_langner]

Hi Marcel,

please post the link to the new threat.

Thanks

Peter

[MarcelRabe]

Hi,

Thanks for the hints. I have a simular problem. The RC 6401 appears in the trace files also but the unescape_url=no as described in sapnote 650521 didn't seem to help. I've added a userid and pwd in the basic authentication part in the preprocessor.ini file and this did the trick.

But now the really strange part. I'm trying to document class SOLARGNSRC (Solution Manager) which holds over 30.000 documents. But only 6300 documents are passed to TREX and after indexing has been completed even those documents cannot be found using the full text search.

Has anybode been succesfull in indexing this document class?

rgds

Marcel Rabe

[peter_langner]

Hi Marcel,

your question should be a new Thread. Are you shure, that the documents were indexed? Have you checked if an index with documents exists on the TREX server?

Peter

[former_member66232]

Hi Peter,

Yes we do have a proxy. I've tried with and without proxy setting and get the same result.

Will also update thread as info becomes available.

[peter_langner]

Hi Paul,

here is the answer for my Problem:

Dear Peter,

unfortunately it seems to be a bug in the TREX installation routine.

Although you didn's specify the proxy server, when installed the TREX

this was set in the TREXPreprocessor.ini:

#proxy

proxyhost = proxy

proxyport = 8080

So I would like to ask you to do the following:

1) Edit the TREXPreprocessor.ini file as:

#proxy

proxyhost =

proxyport =

2) Apply the note 650521 (add the entry unescape_url = no into the

httpclient section of the TREXPreprocessor.ini

3) Restart TREXPreprocessor

4) Reindex documents

Best regards,

Mikhail Sirotkine

TREX Development Support

I hope it solves yout problem too.

Regards,

Peter

Edited by: Peter Langner DEGI on Nov 5, 2008 9:02 AM

Edited by: Peter Langner DEGI on Nov 5, 2008 9:03 AM

[former_member66232]

Hi Peter,

Have you resolved this yet? I have this same error on TREX 7.1 trying to index DMS content via ABAP connection.

Also notice this error in the RRFC trace Log which I think may be the cause but can't find how to correct it. I have a job in OSS but still waiting on it!

Table size mismatch for SRET_INDEX_STORAGE_SEARCHING:DOCATTRTAB, abap=27,rfcserver=29

All other config etc has green lights.

Thanks

Paul

[peter_langner]

Hello Paul,

no, the problem has not been solved so far, I am still working on it. I have also posted and notice to sap. Are you using a proxy? Check it you settings are o.k. (see note 571798). Try to put the contentservers url into the hosts file on your trex server.

As soon as I have a solution, I will put it here.

Regards,

Peter

[frank_friedrich]

Hi Peter,

#proxy

proxyhost = proxy

proxyport = 8080

is in you ini-file.

This means that you are using as a proxy for TREX the proxy server "proxy" and the proxy port 8080.

When you want to use no proxy the entries must be:

#proxy

proxyhost =

proxyport =

You want to use only the portal for the TREX search?

In this case the communication from portal to TREX is over the nameserver via TCP/IP.

The mentioned port in you trace must be accessible (check firewall rules).

After the name server has send back the trex http server name and port all the other portal trex communication will be used the trex http server. And also the preprocessor is using this communication.

Do you have check in the portal the TREX monitoring.

There must be a green traffic light for name server, http server, queue server and index server.

When you will not see all the mentioned server parts of the trex you have a communication problem regarding proxy settings or fire wall rules.

best regards

Frank

[peter_langner]

Hello Frank,

thank you fpr your assistance. I have changed the entry in the ini-file, but it did'nt help. I have even put a comment to the both lines - but it did'nt fix the problem. I habe again checked the ports - they are all open. The nameserver is up and running, but it seams not to be reachable. There is no http server, because I use no portal. I am using TREX in combination with Solution Manager 7.0. I want to use it for the solution database as well as for searching the documents attached to the orders.

If you have any other idea any help is welcome.

Regards

Peter

[frank_friedrich]

Hi Peter,

try to use the URL of the document in the web browser which is running at the TREX server. If the document is open, than compare the proxy settings of the web browser with the proxy settings of the TREX server.

These settings must be the same.

Best regards

Frank

[peter_langner]

Hello Frank,

thank you for this very helpfull hint. I was already asking me, if the proxy settings are a problem. During the installation of TREX we selected "no proxy" and this is also what the internet explorer is showing. On the other hand I found the following entry in the TREXpreprocessor.ini file:

"[httpclient]

use_httpclient =yes

#Check valid mimetype before retrieving whole page

#(otherwise =yes always whole page is retrieved)

always_use_get = no

#timeouts in ms

conntimeout=10000

readtimeout=10000

sendtimeout

continuetimeout

#basic authent

user=

password=

#proxy

proxyhost = proxy

proxyport = 8080

proxyuser =

proxypassword =

max_contentlength = 0

add_attributes = no

reuse_connection= no"

Do you know, if this is the default entry, if you choose "no proxy" during installation?

Do you have any other ideas, what might be the problem?

To me it seems as if it has something to with the reachability of the nameserver. In the TrexTCPClientAlert_s0002000.trc it says:

"[2924] 2008-09-24 14:25:33.943 e TrexNet EndPoint.cpp(00256) : ERROR: failed to open channel 10.25.10.200:30201! reason: connection refused"

and in the TrexRfcServerAlert_s0002000.trc it says:

"[1492] 2008-09-24 14:28:13.283 e TREXRfcSvr rfc_sapretrieval.cpp(00526) : Error sap_internal_index

[1492] 2008-09-24 14:28:13.283 e TREXRfcSvr rfc_sapretrieval.cpp(01016) : rc: 1."

and in TrexAlertServerAlert_s0002000.trc it says:

"[3084] 2008-09-24 14:33:28.072 e TrexNet RfcBufferedIO.cpp(00037) : ABAPModuleInterface::ABAPModuleInterface: RfcGetFunctionInfoAsTable failed with rc: 2 and error message FU_NOT_FOUND"

But the RFC connection is established and with the SAP testing tool I can reach the client from the server and vice versa.

Thank you in advance for your help ,

Peter

[Former Member]

Peter,

You might want to look at this thread.

https://www.sdn.sap.com/irj/sdn/thread?threadID=178454

Good Luck.

Sandeep Tudumu

[peter_langner]

Thank you for the hint. I have checked this thread before. It is a similar but different situation. In my case, the hostname is well known and an access to the documents is given. Also I am not using the portal, but the ABAP stack as a client. So the given path to the URL generator in the portal is not valid in my case.

It must be somthing differend...

Peter