TREX Performance

Hello,
our TREX Server has 3 Ghz, 2 CPUs and 4 GB ram. We have indexed 145.000 Documents (30.000 in Windows Filesystem, 105.000 in Lotus Notes).
A Query on this index takes about 15-20 seconds.
Is this a normal response time? I expected, that TREX is faster.
The Folder "index" on TREX Server has 2,7 GB, but when i start a search, TREX uses only 500-900 MB of his main memory.
It seems as if trex executes one query after the other, when more than one query is executed.
Is there a possibility to tune this, for example to increase the number of processes to execute queries? I have read much about preprocessor tuning, but how can i tune the server, that executes queries (i think this is indexserver).
I read the documents about trex and km tuning, but i couldn't find an answer there.
Thank you
regards
Sven

Hi Sven
Unless you have a very special setup most users dont need to search all indexes, but perhaps only 2-3 indexes. Think about which indexes are pre-checked when a user enters the advanced search window - this can be changes in the search options set you are using.
Don't display the total number of hits in the search result - instead you should use the option to only display 1-10 >> more.... In this way the search should avoid to check permissions for all the document hits. This can also be changed in the search options set.
Permissionscheck is taking A LOT of time when the portal is rendering search results from external datasources (depeding on what kind of integration is done, of course). I believe this is the reason why your search from search.py is so much faster - I dont belive there is any permissionscheck done searching throug search.py (but I'm not sure, though).
Don't create 2-3 big indexes, use smaller indexes and precheck only a few of them (those searched most frequently). You could most likely improve your performance with more slave servers that handle the search requests. Making the indexing server more powerfull will only make indexing perform better. But if you have both functions on the same machine, you are of course interested in upgrading that in order to achieve better performance.
Kind regards,
Martin

Similar Messages

  • TREX Performance Concerns

    Hi all,
    We are interested in using TREX to handle duplicate prevention for business partners using BAS. Should there be any concerns with regards to performance if we have TREX installed on our CRM system?
    Thanks in advance. Points will be awarded for helpful answers.

    Hi,
    Please see TREX Performance Settings.
    Most probably you will need to adjust some settings to optimize performance.
    http://wiki.colar.net/sap_crm5_0_isa_trex_performance_settings
    Regards
    -jemaru

  • Portal & TREX Performance problems!

    Hi All
    I am having an issue with a portal we have just implemented, in terms of performance and TREX processes.
    It appears TREXIndexServer.exe consumes a large amount of processor resource and causes the portal to stop working. When the process is killed off everything goes back to normal!
    Can anyone help me fix this?
    Thanks
    Phil

    We currently only have 7 indexes running on the portal and they run twice a day once at 8.00am and 12:30pm indexing and classifying a total of around 300 documents. These are all stored in the database!
    The Server its self is quite high spec and should not have a problem.
    It seems that the TREX engine !
    Hope this helps!
    Phil

  • TREX Performance Check - HOW TO

    Hi Experts,
    I would like to ask help how can i check if TREX is the cause of my performance issue?
    Where can I check if TREX is the reason why I have performance issue?
    And also how can I check if there is also performance issue with IPC?
    Thank you very much.
    Thanks,\
    -jemaru

    Jemaru,
    Trex monitoring : http://help.sap.com/saphelp_nw04s/helpdata/en/eb/aaa0af02a0734a82c0241f5f62cb67/content.htm
    Good Luck!
    Sandeep Tudumu

  • TREX issues

    Hello gurus,
       I am going to be working on monitoring TREX (used primarily with e-recruiting). There is a TREX server installed up and running and I want to work on monitoring it and troubleshoot any issues that may arise. I wanted to ask you all to help me out with this. Do you have any recommendations on how to keep the TREX performance high and monitor it for any problems and troubleshoot the issues that may arise...
    Is there any guide for this and if there is one, please point me to the guide and also please let me know your suggestions and recommendations for monitoring the performance of the TREX. I basically am looking in the point of how to monitor the indexes.
    I am also getting an error "Index is temporarily not available. Index = ses: xx_xxxx_xx_xxx_ne error 2972 index state idle. I am seeing this error in index admin. In Index landscape, I am seeing a lot of corrupt indexes.
    Could anybody please help me fix this issue for the client?
    Any suggestions/recommendations?
    Thanks much for all your help in advance,

    Hello,
       Our TREX is with E-recruiting and not the portal. I would like to know where is the place to monitor indexes if the trex is connected to the e-recruiting? If it is portal, we go to the system admin -> system config etc.... I do not know where is the trex index monitoriing if it is erec... Could you please let me know?
    Also, is there a way to turn on the master index server and turn off the slave index server?
    Thanks gurus,
    MK

  • Problem with infotype0000

    Hello everyone,
    I am having a problem while uploading files through bdc for infotype 0000 .When I upload the bdc data I don't get an error message but when I check it in PA30 I could see that there is a problem showing "No data stored for actions in the selected period".So can you please explain why do I get this even when I have uploaded the data.
    Thanks
    Regards
    Navratan

    Hello Jagan,
    Thanks for your reply.Here is the recording generated.
    perform bdc_dynpro      using 'SAPMP50A' '1000'.
    perform bdc_field       using 'BDC_OKCODE'
                                   '/00'.
    *perform bdc_field       using 'RP50G-PERNR'
    *                              wa_itab-pernr .         "'3'.
    perform bdc_field       using 'RP50G-TIMR6'
                                   'X'.
    perform bdc_field       using 'BDC_CURSOR'
                                   'RP50G-CHOIC'.
    perform bdc_field       using 'RP50G-CHOIC'
                                 '0000'.
    perform bdc_dynpro      using 'SAPMP50A' '1000'.
    perform bdc_field       using 'BDC_CURSOR'
                                   'RP50G-PERNR'.
    perform bdc_field       using 'BDC_OKCODE'
                                   '=INS'.
    *perform bdc_field       using 'RP50G-PERNR'
    *                              wa_itab-pernr .      "'3'.
    perform bdc_field       using 'RP50G-TIMR6'
                                   'X'.
    perform bdc_field       using 'RP50G-CHOIC'
                                   'Actions'.
    perform bdc_dynpro      using 'SAPLRHWH' '0400'.
    perform bdc_field       using 'BDC_OKCODE'
                                   '=TREX'.
    perform bdc_dynpro      using 'MP000000' '2000'.
    perform bdc_field       using 'BDC_CURSOR'
                                   'T528T-PLSTX'.
    perform bdc_field       using 'BDC_OKCODE'
                                   '=UPD'.
    perform bdc_field       using 'P0000-BEGDA'
                                   wa_itab-begda .     "'01.04.2015'.
    perform bdc_field       using 'P0000-MASSN'
                                   wa_itab-massn.       "'01'.
    perform bdc_field       using 'P0000-MASSG'
                                   wa_itab-massg  .    "'01'.
    perform bdc_field       using 'PSPAR-PLANS'
                                   wa_itab-plans.     "'99999999'.
    perform bdc_field       using 'PSPAR-WERKS'
                                   wa_itab-werks .     "'1234'.
    perform bdc_field       using 'PSPAR-PERSG'
                                   wa_itab-persg .  "'B'.
    perform bdc_field       using 'PSPAR-PERSK'
                                   wa_itab-persk.    "'B2'.
    perform bdc_dynpro      using 'MP000200' '2040'.
    perform bdc_field       using 'BDC_OKCODE'
                                   '/EBCK'.
    perform bdc_field       using 'BDC_CURSOR'
                                   'P0002-BEGDA'.
    Regards
    Navratan

  • Does BIA Blade servers suppot virtualization

    Hi,
    I have a question, would like to know if the BIA Blade servers support virtulization? Is it possible to implement BIA on the Blade servers with VMware installed on it, there by enabling virtualization?. Does the BIA blade servers support virtualization?
    Thanks and Regards,
    Archana

    Hi Archana,
    as Vitaliy already mentioned VMware is currently not supported. The reason therefore is relative easy. BWA is based on the TREX search engine and there is some interesting note which you should take a look (1303814 - TREX 7.1: Usage of TREX on Virtual Machines (VM))
    The two reasons are why it also won´t be supported in near future:
    1) TREX can be used in virtual environments for low-data-volume scenarios. Please note that the published sizing rules may become invalid. SAP has measured TREX performance losses up to 30%, depending on the virtualization solution. If TREX performance issues in virtual server environments arise, make sure you have configured your virtual machine correctly. Reserving memory (and in some cases CPU) are recommended to rule out shortness of resources on the virtualization host in case of resource contention. A detailed resource planning is absolutely mandatory!
    2) TREX should be run on physical servers in high-load environments.
    There is no note especially for BWA, but it´s not supported in cause of the low CPU load of the vmware HW. You won´t get support if you install your prod environment on VMware! If you take the chance you can setup your test environment on VMware, but normally you also won´t get support in case of an error. I think this depends on the philosophy of your company...
    Regards,
    Jens

  • TREX indexing on content server performance

    Hi guys,
    Our Portal is integrated with SAP CRM (using webdav) that manages documents stored in SAP content server. We use TREX to index these documents such that users in Portal can search for these documents. Currently we're evaluating the performance of indexing and searching, thus if we have a heavy load of documents to index, would it affect the SAP CRM/Content server that is the document repository? (such as memory consumption, performance, etc..)
    Thanks,
    ZM

    Hi Chris,
    do you use the ContentServer in the DMS application? If yes, you need to index documents stored in the DMS_PCD1 docu category.
    Regards,
    Mikhail

  • Slow performance after implementing TREX in CRM

    Hi All,
    We have recently implemented TREX in our CRM 7.0 system and from that time we are getting performance issue in two things:
    u2022Saving a New Contact. > Previously it use to take 15 seconds to 1 minute, but now itu2019s taking minimum of 2minutes to 5min maximum.
    u2022Creating Marketing Attributes in a New Contact > Itu2019s taking lot of time. As previously we use to get the option in seconds after selecting the category under > Attribute Set or Attribute Field to get the >u201CValueu201D.
    As we are facing this issue after TREX implementation so we expect that because of TREX we are facing the issue.
    Please suggest root cause of this error and how can we resolve it , Is there any paramter we need to tune after TREX to make the system working fine or any OSS note which contains information for parameter tuning after TREX implementation.
    Please suggest.
    Shivam

    Hi Shivam,
    This is an old post, just wondering how you solve this issue? Do you have any documentation?
    Thank you.
    Regards,
    Edison Wang.

  • How to configure one TREX host with multiple index servers ?

    Hi All,
    Does anyone know how to configure TREX on the one host,
    with multiple index servers ?
    Reason for this is to make better use of resources available on the host server(4 Gig, 4 Processor, Windows2003), to improve the search performance of
    our KM content for portal users.
    I am using TREX 7 and have not been able to do this,
    despite reading the Single and Distributed install
    documentation.
    Any help would be appreciated.
    Regards,
    Andres

    Hi Andres,
    To make use of the RAM a Server provides you have to run two indexserver processes (each can then consume 2 GB);
    Proceed like this:
    1. Go to TREXdeamon.ini; check if section [indexserver2] is there (it is already provided, but not active in standard installation)
    2. In TREXdeamon.ini go to
    [daemon]
    references sections below
    programs=nameserver,preprocessor1,indexserver1,queueserver,alertserver
    and add indexserver2 here. Restart TREX; second porcess is then started; can be checked in TREX monitor in Portal as well
    3. To distribute existing indexes to the new process, start TREXadmintool and go to Index: Landscape
    Go to the last two columns and move the indexes (move master here/secondary mouse click)
    If you don't distribute the indexes the new index server process will be regarded when an new index is created.
    Hope this helps!
    cheers
    Bettina

  • TREX - Configuring Distributed Slave with Decentralized Data Storage

    I am creating a distributed TREX environment with decentralized data storage with 3 hosts.  The environment is running TREX 7.10 Rev 14 on Windows 2003 x64.  These are the hosts:
    Server 01p: 1st Master NameServer, Master Index Server, Master Queue Server
    Server 02p: 2nd Master NameServer, Slave Index Server
    Server 03p: Slave NameServer, Slave Index Server (GOAL; Not there yet)
    The first and second hosts are properly set up, with the first host creating the index and replicating the snapshot to the slave index server for searching.  The third host is added to the landscape.  When I attempt to change the role of the third host to be a slave for the Master IS and run a check on the landscape, I receive the following errors:
    check...
    wsaphptd03p: file error on 'wsaphptd03p:e:\usr\sap\HPT\TRX00\_test_file_wsaphptd02p_: The system cannot find the file specified'
    wsaphptd02p: file error on 'wsaphptd02p:e:\usr\sap\HPT\TRX00\_test_file_wsaphptd03p_: The system cannot find the file specified'
    slaves: select 'Use Central Storage' for shared slaves on central storage or change base path to non shared location
    The installs were all performed in the same with, with storage on the "E:" drive using a local install on the stand-alone installation as described in the TREX71InstallMultipleHosts and TREX71INstallSingleHosts guides provided.
    Does anybody know what I should try to do to resolve this issue to add the third host to my TREX distributed landscape?  There really weren't any documents that gave more information besides the install documents.
    Thanks for any help.

    A ticket was opened with SAP customer support.  The response to that ticket is below:
    Many thanks for the connection. We found out, that the error message is wrong. It can be ignored, if you press 'Shift' and button 'Deploy' (TREXAdmin tool -> Landscape Configuration).  We will fix this error in the next Revision (Revision 25) for TREX 7.1.

  • TREX/Content Server: Full text searches not working

    I've setup TREX and Content Server in a sandbox environment and have been trying to get my head around DMS.  I've completed what I've believed are the required steps (outlined below) and I have submitted 12 documents of various types into the respository.  I'm able to perform meta data searches but not full text searched.  Any ideas what I might have missed?
    Much thanks!!!!
    /Greg
    0. Defined connection to TREX Server: TREX_42 (SM59); Using TREXADMIN all services are green.
    1. "Define Document Area" - DMS is present; assumed I would use this one.
    2. "Defined Indexing Readiness" - selected "Index Release" & "Classify" for DMS; set "Search Srv" to DMS_TEST
    3. "Define Content Respository" - Created Z_CR_001 and maintained it as follows:
      - Documentation Area: Document Management System
      - Storage Type: HTTP content server
      - HTTP Server:  srv-sptxd2-dal.intervoice.int  (content server)
      - Port: 1090
      - HTTP Script: ContentServer/ContentServer.dll
      - Physical Path: /usr/sap/IRD/SYS/global/
    4. "Define Content Categories" - Created ZCSSTRUCT; maintained as follows:
      - Document Area: DMS
      - Content Rep: Z_CR_001
    5. Checked CSADMIN for Z_CR_001 - All Green
    6. Went into SKPR07 (monitoring tab); entered DMS_PCD1 and "Indexing possible", "Auto indexing" and "DocArea activated" is all green.
    7. Went into SMRO
      - created SSR: "Search Server ID" = DMS_TEST, "Search engine" = DRFUZZY
      - Verified "RFC Destinations"; passed tests; all green.
    8. Went into SKPR07; using DMS_PCD1 as document class. 
      - executed "Clean up table entries"
      - deselected "Limited to selected language"
      - executed reindex. "Copied successfully from search engine"
      - Selected the created index category
      - Set the "Document area" indicator
      - execute "Trigger process"; 12 documents referenced
    9.  Going into TREXADMIN I can see the created index; however the document count = 0.  Also there is no queue created -- should there be?
    Edited by: Greg Preston on Aug 5, 2009 9:11 PM

    Dear
    this link will help you,
    TREX Architecture
    http://help.sap.com/saphelp_erp60_sp/helpdata/EN/71/b26b2bfdc4eb47ab5432e8444290ce/content.htm
    File Formats Supported by TREX
    http://help.sap.com/saphelp_erp60_sp/helpdata/EN/55/cb634114b78047e10000000a1550b0/content.htm
    Detail about TREX Configuration  (Basic to advance)
    http://help.sap.com/saphelp_erp60_sp/helpdata/EN/46/bab1d48b0a1514e10000000a114a6b/content.htm
    Regards
    Tushar Dave
    Edited by: Tushar Dave on Aug 6, 2009 9:59 AM

  • How to use TREX in Netweaver Portals

    Hi,
    I am new to TREX and i installed TREX 7.0 with Netweaver 2004's java edition and performed Post Configuration for Java client HTTP settings.
    Now i am unable how to use TREX in Portals for searching the documents in KM.
    Can any one help me how to use the TREX for searching the documents in Km if possible plz attach the screen shots.
    ADV Thanks
    Regds
    Phani

    Hi Phani,
    You can test your TREX server by executing the following step:
    Log on to your TREX server and open up the TREX console which you can find in the TREX program folder. Normally you are now inside the python_support folder.
    Execute the following script by typing: "python runInstallationText.py" and hit Enter.
    The script carries out the following tests:
    ● Deleting any test indexes that were generated for a previous script run
    ● Creating a test index
    ● Indexing documents
    ● Testing search functions
    ● Exact, error-tolerant (fuzzy), and linguistic searches
    ● Search using Boolean operators such as AND and OR
    The results are displayed at the end of the script run. You see the tested calls and their statuses (OK or Failed).
    When you run the script for the first time, the call “Delete Index” has the status Failed. This is because there was no existing text index to be deleted. If this is the only cell with the status Failed, the test was successful.
    Below you will see the final output generated by the test script. As stated near the bottom of this screenshot our TREX is running very well or not if there is a problem.
    Best regards,
    Tom.
    PS. Don't forget to reward helpfull answers

  • Trex search E-recruitment

    Hello all,
    I have a question regarding a service in an abap engine:
    I have configured a trex engine for e-recruitment and I am getting an error code 8002.
    if it happens in a java engine I now that you have to configure a java service called URL Generator service, in that service you have to put the hostname. is there a similar thing in abap?
    did anyone encounter the problem?
    Thanks
    Asher

    The problem could be that the pre-processor loads the documents and performs preliminary parsing/analysis on the document before the index server can get to work on it. The URL generated will be used to load the documents and must have sufficient information else the pre-processor fails.
    You might need to have a look at the following links for possible soultions depending on your situation:
    /thread/346637 [original link is broken]
    http://help.sap.com/saphelp_nw04/helpdata/en/7d/236cfa17034a37a439dc392ec59eb0/frameset.htm

  • TREX Engine not working in PI

    Hi Experts ,
      My team has configured TREX with PI, to monitor PI messages.
         I can see that , unlike the Integration Engine which has about 190000 messages,  there were no messages indexed for Adapter Engine, is this expected ?
      Am I missing some important configuration / settings.
    Please suggest.
    Regards,
    Arnab.

    Hi Arnab,
    For PI 7.1 with EHP1, then you can following threads to search the messages using Trex:
    http://www.sdn.sap.com/irj/scn/index?rid=/library/uuid/d05c14c3-34af-2b10-a1b1-fa2a39e0d2ae
    /people/niki.scaglione2/blog/2010/02/22/payload-searching-without-trex
    It should be there, sicne Trex server is installed.
    Else, you have to write the ABAP code for search:
    /people/sravya.talanki2/blog/2006/02/21/abap-based-trex-in-xi-proto
    Also, if you have configured correctly cross check the steps below:
    The global settings is valid for all components u have selected for indexing. You can use the default value itself.
    ther are parameters like
    1) Maximum number of messages per package to be transferred to the TREX server
    2) Maximum size in KB of message packages transferred to the TREX server
    3) Time interval in minutes for the periodic transfer of message packages to the TREX server for indexing
    4) Time interval in hours for the periodic reorganization of the index of a component on the TREX server
    5) Time period in days for the retention period of messages in the index
    6) Time interval in minutes for the periodic processing of the TREX queue
    7) Maximum number of entries for processing in the TREX queue
    Also ther are filters available. if the filter is empty then all the messages will get indexed and will be available in search,
    if u enter any details in filter option, then that particular interface only be indexed and only indexed messages will be available for content base search. The use of filter will improve the performance in case of large message flow and also it will save your hardware resources. but make sure that all required interfaces will be available in filter or other wise will searching you will get confused.
    Also, check sap note: 1266924 if everything is fine!
    Hope this helps.
    Reg,
    Sushama
    Edited by: sushama pandey on Oct 5, 2010 6:53 AM

Maybe you are looking for