2 Replies Latest reply: Apr 6, 2016 7:31 AM by Jonas Gros RSS

    Corrupt Meta File

    Brent Nichol

      Recently, I had a .Meta file become corrupt and brought my QlikView Server to its knees.

       

      The following lines were logged dozens of time per second, adding up to 2.2 million rows in a 30 minute period...

       

      ErrorShared - Key.ID: Error reading shared objects of type , Id: , StopLoadingWhenSharedFileIsCorrupt: 0, Document: DocumentName.qvw.Meta
      ErrorShared - Key.Type: Error reading shared objects of type , StopLoadingWhenSharedFileIsCorrupt: 0, Document: DocumentName.qvw.Meta
      ErrorShared - Key.User: Error reading shared objects of type , user: , StopLoadingWhenSharedFileIsCorrupt: 0, Document: DocumentName.qvw.Meta

       

      Not until I was able to delete the corrupt Meta file and restart each server did the QlikView Server begin responding.

       

      Would the QVS have eventually handled this automatically? 

       

      Is there a way to prevent this from reoccurring?

       

      Regards,

      B

        • Re: Corrupt Meta File
          Paul Verkooijen

          Hi bnichol,

           

          Are you sure it was the .Meta file that was causing the issue, or have you also removed the .shared file ?

          I've have never seen a .Meta file with errors, however a .Shared file containing objects by deleted (non existing) users causes issues.

           

          In the reference manual is a description how you could automatically scan and repair the Shared files.

          It is called the "Shared file Cleaning Tool"

          • Re: Corrupt Meta File
            Jonas Gros

            Same problem today for me. Corrupt Meta-File killed the whole Server. Nothing helped until stopping server-service and delete the meta-file. Start the service again and enjoy! No applications could be opened up. Nothing was responding on management console. Just the eventlog helped me out.

             

            I hope this never happens again. 2 hours downtime. First bigger downtime since 5 years.