2 Replies Latest reply: Feb 3, 2014 5:02 AM by Peter Cammaert RSS

    How to increase QV server based speed?



      I have QV based on server with ca 3-5 million data rows uploaded and QV based on a server. Using QV via web-browser with that kind of data volume is really slow. The lag is big. Can you recommend any smart ways to increase the speed of web-browser based version?

      - I'm using Explorer. I was told that Chrome is much quicker but for some reason I'm having problems with logging in (Explorer works Chrome doesn't - I all the time get "login failed" messages)


      Many thanks,



        • Re: How to increase QV server based speed?
          Michael Laenen

          How does look like you data model?

          How is you server ? RAM/CPU/ Nbr of concurrents User? Size of Document?

          How is the content of the complexity of the document? (ex displaying straight table with complex set analysis on 5 milions rows is CPU consumming, add some aggr function within and you'll need also some more RAM...

          Is the QVS running on a virtual machine, if yes does the VM is dedicated to QV?

          You could run some load testing with JMeter



          • Re: How to increase QV server based speed?
            Peter Cammaert

            5 million rows should not be any problem for QlikView. Probably more of a problem because of one or more of the following:

            • Not enough memory. Monitor your server memory usage. QVS becomes awfully slow when running short on memory.
            • Not enough CPUs for parallel resolving of selections and associations. Monitor your server CPU usage.
            • Your datamodel may be sub-optimal. Try not to use any excessive table chains. Throwing it all in one big table makes for a incredibly responsive application, on condition that the previous items are all ok.
            • Chrome indeed has an excellent javascript engine, and performance improvements are substantial when switching over from IE. I suggest you start a separate thread to resolve your "login failed" problem.