4 Replies Latest reply on Apr 4, 2017 2:40 AM by Data Central

    AMAZON EMR Hive issue

    julien PHILIPPE
    julien PHILIPPE Mar 21, 2017 9:41 AM

      Hi, I have :

      - a PRD tableau server 9.3

      - a TST tableau server upgraded from 9.3 to 10.2

       

      Both are connecting to the same datasources, but the connection is failing from TST tableau server 10.2 with the following error:

       

      [Amazon][ODBC] (10430) Not enough information provided to establish a connection to the data source and specified to not prompt for more information. Missing required settings: {[PWD] [UID]}

       

      The driver is 'Amazon EMR hive ODBC driver' on both TST and PRD, the server is a windows NT6.2, the connection is working from tableau desktop installed on the TST server !!!

       

      tableau.jpg

      • 437 Views
      • Categories: Upgrading
      • Tags:
        • 1. Re: AMAZON EMR Hive issue
          Russell Christopher
          Russell Christopher Mar 22, 2017 6:57 AM (in response to julien PHILIPPE)

          I suggest you attempt to narrow the  issue a bit:

           

          • What happens when you create a new data source with Tableau Desktop and then publish it to Tableau Server (TST). Does the new one work?
          • Have you downloaded the latest EMR drivers from the Tableau/Amazon website and installed them on TST? Driver requirements do change from time to time. I don't know if EMR's have, but I'd take a look at that.
          • 2. Re: AMAZON EMR Hive issue
            julien PHILIPPE
            julien PHILIPPE Mar 22, 2017 8:18 AM (in response to Russell Christopher)
            • What happens when you create a new data source with Tableau Desktop and then publish it to Tableau Server (TST). Does the new one work?

            --> yes we did that test, no it didn't worked.

            • Have you downloaded the latest EMR drivers from the Tableau/Amazon website and installed them on TST? Driver requirements do change from time to time. I don't know if EMR's have, but I'd take a look at that.

            --> yes I tried to update to the latest EMR driver (tried both 32 and 64)

            • 3. Re: AMAZON EMR Hive issue
              Russell Christopher
              Russell Christopher Mar 22, 2017 10:10 AM (in response to julien PHILIPPE)

              Interesting.

               

              Since you've validated that the ODBC driver itself works on TST (via your Desktop test), we can rule driver out now.

               

              One thing that I've seen from time to time is a situation in which Tableau Server (generally the "RunAs" user) doesn't have permissions on the folder / profile information associated with the ODBC driver. If Server doesn't have permissions on the file system to get at the driver, you'll get a failure (although I recall the message was different than what you're showing here).

               

              In your shoes, I'd do the following:

               

              • Get a support case open if you haven't already - I'm not a tech support guy, I just play one on TV now and then
              • Try logging into the console AS the same user Tableau Server is running under and verify if you have permissions on the driver and that you can still connect...
              • AND/OR: Temporarily change the RunAs user that Tableau Server is running under to a local ADMIN who "should" have file system access to everything on the machine - see if the behavior changes
              • AND/OR: Change file permissions on the drivers's folder, assigning full control to "Everyone". You might need to bounce Tableau too in order to force the RunAs user to log back in and get the "new" permissions. Not sure.
              • 4. Re: AMAZON EMR Hive issue
                Data Central
                Data Central Apr 4, 2017 2:40 AM (in response to Russell Christopher)

                Hi, I created a support case as you suggested. The connection works when we republish from a 10.2 desktop (and not 9.3 desktop). It's a big issue for a production migration with a lot of data sources...