3 Replies Latest reply on Jun 7, 2012 3:55 PM by 940688

    Large Query quits after generating results in workspace

    940688
      We have a large query that generates about 3.5 million rows in IR 11.1.2.1. The results are then used in table sections in the same bqy. When we publish this as a job in workspace, it seems to fetch all the rows and quits after that(before getting to the bqy table sections) with the following error in the '1_default1_irjob_a832ea59b6fab72bS66b45d10133eb2ce195S777b_stdout.log':
      Process reached 1536 MB memory limit….starting shutdown sequence.

      Can anybody elaborate what memory limit the message is talking about?Is this just temp space in the Hyperion server(specific for workspace jobs)?If so, can we increase the memory/where do we do this? Is there a limit to the number of rows that IR jobs can query and bring in?
      When we run the query from IR studio(desktop), it runs fine.

      Thanks.
        • 1. Re: Large Query quits after generating results in workspace
          Vivek Chetiaparath
          It is taking about the memory limit for IR job service which is 1536.That is the memory limit it can handle.if anything goes beyond ,the query will fail.
          • 2. Re: Large Query quits after generating results in workspace
            WVanSluys
            a couiple of issues/questions here.

            Do you really have a report that shows 3.5 million rows of data on it? That is a bit extreme for an Analysis/Report

            how many columns wide is the data? how big is the file on the desktop uncompressed with all the data? are you doing sorts or computed columns in the BQY on the data? How many Chart, Pivot, Report sections

            What is your EPM install 32bit or 64 bit?

            You might want to look at:
            - Aggregating data in Query or Database.
            - Using a different Tool. (Production Reporting (SQR)
            - if using 64bit increase the JVM settings (min/max) for BI service.
            • 3. Re: Large Query quits after generating results in workspace
              940688
              The real issue seems to be when we carry over the data from the results to bqy table sections. We have since split up the query with various limits, and are writing to Oracle tables instead, and then pulling it from Oracle into the bqy for our analysis. We had already played with the JVM settings in our 64-bit system.
              We will be moving towards OBIEE pretty soon here.Just wondering if it is any different from IR in handling huge volumes of data.
              Thanks for your insight.

              Edited by: 937685 on Jun 7, 2012 8:54 AM