1 2 Previous Next 28 Replies Latest reply: Apr 27, 2012 8:07 AM by 920613 Go to original post RSS
      • 15. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
        user5810354
        Also Have a look at metalink article 189819.1
        • 16. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
          920613
          Hi, I do not see those values being available from the Repository Manager applet, configure button. Is there are file where I could see these manually?
          • 17. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
            Srinath Menon-Oracle
            Hi ,

            ORA-01013: user requested cancel of current operation

            This error indicates that the DB Connections are not enough for UCM indexer process / search to complete .

            Usually in these kind of cases I would bounce DB server completely - service and listener and then test the issue .

            Along with that the Optimize feature is something that you should check for first . It should be executed first , DB server restarted and then test .

            Hope this helps .

            Thanks
            Srinath
            • 18. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
              don ray
              If you are storing the Search Index in the same database as the Content Server Database (CSDS Data Source in WebLogic and SystemDatabase in UCM Providers) I would add both of the following to the config.cfg file:

              SearchIndexerEngineName=OracleTextSearch
              IndexerDatabaseProviderName=SystemDatabase

              We were also having similar problems and made the following changes tot he database:

              memory_target=805306000
              shared_pool_size=250000000

              One other thing we did that may have also helped was that we increased the JAVA Heap size for the Content Server via WebLogic Administration Console. Open the WL Admin Console, expand Environment, select Servers, Select UCM_server1, Select the Server Start tab and enter the following in the Arguments box:

              -Xms512m -Xmx2048m -XX:CompileThreshold=8000 -XX:PermSize=256m -XX:MaxPermSize=1024m

              You may alos have to create an additional Database Provider in UCM for the OTS search (we had to but we have two databases, one for the Content Server and one for the Search Index). When you create the Provider you will check the box for Use Data Source then put CSDS in the box for the data source. Leave everything as the default not JDBC Driver or Connection String are needed. Also anything below the Test Query box is also not needed so just leave it as the default. All of these values are picked up by the Data Source in WebLogic.

              We can process 100,000 Documents with 800,000 of them Full Texet Indexed in 3 Hours. It was taking longer, but after the changes to the Database I believe that helped as well.

              Restart and rebuild the index. Also, is this a single server or is it a multiple server environment?
              • 19. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
                920613
                Hi,

                This second time our indexer took 7 days to index 108,000 documents. After indexing, when I perform a search we're getting the following error:

                Event generated by user 'test_user' at host 'hostname'. Error occurred while processing. Unable to return results. [ Details ]
                An error has occurred. The stack trace below shows more information.

                !csUserEventMessage,configuration_manager,qawc.doe.gov!$!csSearchUnableToReturnResults
                intradoc.common.ServiceException:
                *ScriptStack GET_SEARCH_RESULTS
                3:getSearchResults,**no captured values**

                I have the following entries in my config.cfg:

                SearchIndexerEngineName=OracleTextSearch
                IndexerDatabaseProviderName=SystemDatabase
                AdditionalEscapeChar=_:#
                • 20. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
                  user5810354
                  Just another update on this thread.

                  The error involved here is not actually the indexer failing but infact the the optise job failing.

                  When you do a full index rebuilt using OTS the optimise job runs at the end to clean the index tables up. this has a timeout out built into it, and if you pass this timeout, you get this error, plus a schedule job set up to run every hour (if you view your logs you should see this error every hour).

                  If you run the optimise job manually there is no timeout and therefore completes.

                  Turn the scheduled job off!!

                  Oracle have a bug raised to be able to set the timeout!
                  • 21. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
                    920613
                    Hi All,

                    Here is the latest update on my Oracle Text Search indexing issues.

                    1. I was able to successfully configure and index using OTS on our sandbox environment with 1,500 documents. The sand box is a single node and no clusters.

                    2. I used the same configurations on our DEV and QA enviroments with 108,000 documents and I'm running into severe indexing performance issues. Both QA and DEV are clustered with 2 nodes.

                    QA indexer is indexing an average of 18,000 documents per day...total rebuild takes about week to week & half.

                    DEV indexer is awful and indexed only 1956 documents in 16 hrs and at this rate, I'm afraid it will take forever to finish the indexing rebuild.

                    3. My database server settings on DEV and QA are upped to :

                    Max SGA size 2,137,886,720 > 10,689,474,560
                    Shared Pool 822,083,584 -> 1,040,187,392
                    Buffer Cache 520,093,696 -> 7,415,529,472

                    4. My Java settings in the UCM server start up are:

                    -Xms512m -Xmx2048m -XX:CompileThreshold=8000 -XX:PermSize=256m -XX:MaxPermSize=1024m

                    5. Our DBAs ran a trace and has the below comments:

                    Ran a trace on the process overnight on the database since the information you provided is leaning towards database performance. The process ran overnight before we turned off the trace and found 99% of the time was waiting for a response.

                    For the database SQLNETmessage from client is the equivalent to the time in sqlplus between you typing the commands. Like if you type select * from users; and it gives the response … then you are working on something else for 10 minutes and then type select * from projects.

                    While the database only works for a second the trace time was 10 minutes and 1 second.

                    We do not have any errors in the Content Server and Database logs via the Administration link in the Content Server.

                    Any thoughts on what could be impacting our indexing speed ?


                    Thanks!
                    • 22. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
                      Srinath Menon-Oracle
                      Hi ,

                      Indexer performance to an extent can get downgraded due to the FT indexes being un-optimized .

                      To alleviate that problem check to see what is the extent of fragmentation happened on the indexes .

                      DBA can run this before and after the optimize_index runs. In the report there's a part about "fragmentation statistics." It will show a percentage of fragmentation. After the optimize runs it should show 0% or near zero.


                      How to run the index_stats report to view fragmentation of the index:

                      --Run the index_stats report to see some statictics.
                      --This section of code creates a table called output to store the stats report.
                      drop table output;
                      create table output (result CLOB);

                      declare
                      x clob := null;
                      begin
                      ctx_report.index_stats('FT_IDCTEXT2',x); (or IDCTEXT1 for OTS)
                      insert into output values (x);
                      commit;
                      dbms_lob.freetemporary(x);
                      end;
                      /
                      --Print out the index_stats report.
                      --Run this in SQL Developer as statement (Ctrl-Enter) not as a script (F5)
                      select * from output;

                      --Double click on the output line to see the full report.

                      DBA should run the below job on the fulltext index, and also schedule this job to run weekly from now on (during weekends or nights). The job will not be noticed by users, but will defragment the full text search index. This job is necessary to keep the fulltext index in top performance, since fragmentation in the index can significantly degrade performance on searches.

                      For DB FT :

                      begin
                      ctx_ddl.optimize_index('FT_IDCCOLL1','FULL', parallel_degree =>'1');
                      end;

                      -------------------------------------------------------------------------------------

                      For OTS :

                      begin
                      ctx_ddl.optimize_index('FT_IDCTEXT1','FULL', parallel_degree =>'1');
                      end;



                      Infact for high volume systems this should be run on a daily basis .


                      To make sure if the above operations are to be run on IDCCOLL1 , IDCCOLL2 , IDCTEXT1 ,IDCTEXT2 navigate to UCM Administration - Configuration Information - Active Index .

                      What ever the value is seen there will be the one that will need to be put in the above procs .

                      Thanks
                      Srinath
                      • 23. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
                        920613
                        Srinath,

                        Thank you so much...this resolved my issue.The indexing finished within 3 days and 18 hrs. For the first 3 days the process indexed “45,954 documents” and after running the optimization scripts....in 17 hrs it indexed 62,677 documents. Now, I’m pretty confident that I can re-run the entire process all over in 24 – 48 hrs.
                        • 24. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
                          920613
                          Srinath,

                          I'm having another issue now and wondering if you could share some thoughts on this? I'm seeing a different error in my content server and I believe this is OTS related error.

                          Unable to find parsed elements with engine label 'DATABASE.METADATA.ALLDOCS', field name 'xClbraUserList', operator name 'zonehasasword'.

                          Issue Title : Unable to find parsed elements with engine label 'DATABASE.METADATA.ALLDOCS', field name 'xClbraUserList', operator name 'zonehasasword'.

                          Error     Event generated by user 'username' at host 'CIS'. Unable to find parsed elements with engine label 'DATABASE.METADATA.ALLDOCS', field name 'xClbraUserList', operator name 'zonehasasword'. [ Details ]
                          An error has occurred. The stack trace below shows more information.

                          !csUserEventMessage,imtester,CIS!$!csUnableToFindParsedElements,DATABASE.METADATA.ALLDOCS,xClbraUserList,zonehasasword
                          intradoc.common.ServiceException:
                          *ScriptStack GET_SEARCH_RESULTS
                          3:getSearchResults,**no captured values**
                          at recordsmanagement.RmaSearchService.getSearchResults(RmaSearchService.java:115)
                          at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
                          at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
                          at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
                          at java.lang.reflect.Method.invoke(Method.java:597)
                          at intradoc.common.IdcMethodHolder.invokeMethod(IdcMethodHolder.java:87)
                          at intradoc.common.ClassHelperUtils.executeMethodEx(ClassHelperUtils.java:310)
                          at intradoc.common.ClassHelperUtils.executeMethod(ClassHelperUtils.java:295)
                          at intradoc.server.Service.doCodeEx(Service.java:551)
                          at reportpublisher.RPServiceHandler.getSearchResults(RPServiceHandler.java:181)
                          at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
                          at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
                          at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
                          at java.lang.reflect.Method.invoke(Method.java:597)
                          at intradoc.common.IdcMethodHolder.invokeMethod(IdcMethodHolder.java:87)
                          at intradoc.common.ClassHelperUtils.executeMethodReportStatus(ClassHelperUtils.java:324)
                          at intradoc.server.ServiceHandler.executeAction(ServiceHandler.java:79)
                          at intradoc.server.Service.doCodeEx(Service.java:533)
                          at intradoc.server.Service.doCode(Service.java:506)
                          at intradoc.server.ServiceRequestImplementor.doAction(ServiceRequestImplementor.java:1643)
                          at intradoc.server.Service.doAction(Service.java:477)
                          at intradoc.server.ServiceRequestImplementor.doActions(ServiceRequestImplementor.java:1458)
                          at intradoc.server.Service.doActions(Service.java:473)
                          at intradoc.server.ServiceRequestImplementor.executeActions(ServiceRequestImplementor.java:1391)
                          at intradoc.server.Service.executeActions(Service.java:458)
                          at intradoc.server.ServiceRequestImplementor.doRequest(ServiceRequestImplementor.java:737)
                          at intradoc.server.Service.doRequest(Service.java:1890)
                          at intradoc.server.ServiceManager.processCommand(ServiceManager.java:435)
                          at intradoc.server.IdcServerThread.processRequest(IdcServerThread.java:265)
                          at intradoc.server.IdcServerThread.run(IdcServerThread.java:160)
                          at weblogic.work.SelfTuningWorkManagerImpl$WorkAdapterImpl.run(SelfTuningWorkManagerImpl.java:528)
                          at weblogic.work.ExecuteThread.execute(ExecuteThread.java:209)
                          at weblogic.work.ExecuteThread.run(ExecuteThread.java:178)
                          Caused by: intradoc.common.ServiceException:
                          at intradoc.server.DocumentAccessSecurity.appendQueryElement(DocumentAccessSecurity.java:1732)
                          at intradoc.server.DocumentAccessSecurity.buildClauseElementEx(DocumentAccessSecurity.java:1655)
                          at intradoc.server.DocumentAccessSecurity.buildClauseElement(DocumentAccessSecurity.java:1603)
                          at intradoc.server.DocumentAc

                          Edited by: 917610 on Apr 25, 2012 6:09 PM
                          • 25. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
                            Srinath Menon-Oracle
                            Hi ,

                            Do you have xClbraUserList as part of Zone field ? You can check this from UCM Administration - Zone Fields .

                            If yes , then remove it from that list and update , then check to see if the issue still persists .

                            For the index resolution all credit should go to Peter Flies who has provided this solution .

                            Thanks
                            Srinath
                            • 26. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
                              920613
                              Srinath,

                              I cannot find the "Zone Fields Configuration" under my UCM Administration. I have two environments and I can see the "Zone Fields Configuration" in DEV and I CANNOT see it in QA.

                              A while ago, I ad this in my config.cfg #ZonedSecurityFields=xClbraUserList,xClbraAliasList ....however, it has been commented out for a long time. Now, how can I confirm if I have any Zoned fields or not in my QA?
                              • 27. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
                                Srinath Menon-Oracle
                                Hi ,

                                On Dev this issue is happening ?

                                Thanks
                                Srinath
                                • 28. Re: Unable to Index Content in UCM 11gR1-11.1.1.5.0 with Oracle Text Search
                                  920613
                                  Srinath,

                                  My dev is messed up so they're rebuilding the entire DEV. How can i fix this in QA? How can I make the Zone field configuration/administration link show up in Administration link in UCM?
                                  1 2 Previous Next