2 Replies Latest reply on Aug 31, 2011 10:21 AM by Steve Ostrovsky

    Ongoing Replication Problems

    Brian Endres

      Lately, our replicated server gives the following two messages in the Archive Server log:

       

      Error: 0x080040267
      Function: hGetErrCode
      Description: Error running query in SpDbOdbc
      State=01000
      Msg=[Microsoft][ODBC SQL Server Driver][DBNETLIB]ConnectionWrite (send()).

       

      Replication of vault "TCOM Vault" failed.
      Cause: IDispatch error #103

       

      I can't find any way to track what these errors codes mean - can anyone help me out here?  Granted, I can just reboot the server whenever this happens but it's getting to be multiple times a week...

       

      Thanks in advance!

        • Re: Ongoing Replication Problems
          C. Vigneswaran

          Hi,

           

          We are also getting same error message in our replication server. Anybody knows solution for this?

           


          Thanks
          Vignesh

          • Re: Ongoing Replication Problems
            Steve Ostrovsky

            Per SolidWorks KB:

             

            If the archive server service fails to correctly communicate with the SQL server to get a list of files that the scheduled replication is set to process, the schedule may fail and the following errors could show in the archive server log:

            --

            Description: Error: 0x080040267

            Function: hGetErrCode

            Description: Error running query in SpDbOdbc

            State=01000

            Msg=[Microsoft][ODBC SQL Server Driver][DBNETLIB]ConnectionWrite (send()).

            Date: 2010-02-08 03:30:00

            Module: ArchiveServerService

            Function:

            Error code:

            Error code general description:

            ---

            Description: Replication of vault "NCA" failed.

            Cause: IDispatch error #103

            Date: 2010-02-08 03:30:00

            Module: ArchiveServerService

            Function: Replicator::mvStartFullReplication

            Error code: E_EDM_DATABASE_LINK_FAILURE

            Error code general description:

            A communication link failure occurred when accessing the database server. The database service might have been restarted or the network might have gone down. Please try the operation again and contact your system administrator if the problem persists.

            ---

            If the above errors are consistently shown in the log with the same timestamp each day, it is most likely caused by incorrectly installed SQL-DMO drivers causing the SQL communication to fail.  If the error is only seen randomly, it’s more likely a temporary communication problem between the archive server and SQL server.

             

            In these cases, manual on-demand replication will work as it only affects the scheduled replication.

             

            To resolve the issue, re-install SQL-DMO by running the “Microsoft SQL Server 2005 backward compatibility" install from the Enterprise setup media (note that it should be installed if connecting to SQL Server 2008 also). The install is found under \Setup\ISSetupPrerequisites\{2A2B0450-D83A-4779-B083-FC7E50A1BB86}\SQLServer2005_BC.msi or can be downloaded from www.microsoft.com.

             

            If installing on a 64-bit system, the install is found under \Setup64\ISSetupPrerequisites\{2A2B0450-D83A-4779-B083-FC7E50A1BB86}\SQLServer2005_BC.msi

             

            Note that if the Modify/Repair/Remove dialog displays when running the above .msi package, select to uninstall the existing package and then re-run the installation to ensure that all drivers become correctly installed and registered.

             

            Once installed, restart the archive server service and it should now process the replication schedule correctly upon next start-up.  To test that the schedule starts up as it should, create a new temporary schedule for a specific folder containing files not yet replicated, and set the interval to * * *  (this should only be used when testing). If the schedule starts processing the files, there should be a log entry similar to: "Starting replication of <vault>\folder from server #1. There are 10 files that needs to be replicated.".   Once confirmed that the test schedule working, it is recommended to remove it.

             

            Also ensure that the archive server can fully establish a SQL connection by running an ODBC test - see KB solution S-042701.

             

            I hope this helps.