Uploaded image for project: 'Red Hat Process Automation Manager'
  1. Red Hat Process Automation Manager
  2. RHPAM-2738

OOME on org.drools.benchmarks.turtle.buildtime.BuildKieBaseFromContainerBenchmark

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Blocker Blocker
    • 7.7.0.GA
    • 7.7.0.GA
    • BRE
    • None
    • CR1
    • CR2
    • +
    • Hide

      I consider disabling the ExternaliseCanonicalModelLambda is not an option because if it can be enabled on customers environments, we can face performance issues there.

      Show
      I consider disabling the ExternaliseCanonicalModelLambda is not an option because if it can be enabled on customers environments, we can face performance issues there.
    • Hide

      Run the benchmark locally with 7.33 branch.

      Show
      Run the benchmark locally with 7.33 branch.

      java.lang.OutOfMemoryError: GC overhead limit exceeded during execution of org.drools.benchmarks.turtle.buildtime.BuildKieBaseFromContainerBenchmark.* when creating the kiebase. You can see the stack trace on [1].

      After disabling ExternaliseCanonicalModelLambda*, the benchmark was able to run normally. The memory consumption is around 2Gb[2], otherwise it goes to the infinite (I tried until 12Gb).

      • I changed the drools code base but it can be disabled programmatically

      [1] jenkins result: https://gist.github.com/hellowdan/c5571952afd929e792a6633b2a7923db
      [2] 2,2Gb (attachment [memory_consumption.jpg]), sometimes 3,3Gb (attachment [memory_consumption2.jpg])

        1. memory_consumption.png
          31 kB
          Daniel Rosa
        2. memory_consumption2.png
          27 kB
          Daniel Rosa

              mfusco@redhat.com Mario Fusco
              drosabrno Daniel Rosa
              Daniel Rosa Daniel Rosa
              Daniel Rosa Daniel Rosa
              Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

                Created:
                Updated:
                Resolved: