Uploaded image for project: 'Blazegraph (by SYSTAP)'
  1. Blazegraph (by SYSTAP)
  2. BLZG-641 Improve load performance
  3. BLZG-1527

Experiment with heap size, GC mode, and nursery settings

    XMLWordPrintable

    Details

    • Type: Sub-task
    • Status: Open
    • Priority: Medium
    • Resolution: Unresolved
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Description

      Scaleout generally runs larger heaps for client services. These CS nodes are responsible for running parsers and buffering large amounts of data on the heap and incrementally writing on the various indices using the IAsynchronousWriteBufferFactory.

      We should explore the parameter space for bulk loading into blazegraph, especially for cases where the system is not answering concurrent queries while loading data. For such cases it is Ok to have long full GC pause times. This means that it is quite viable to have very large heaps and that can reduce the IO burden through a combination of elided writes and ordered writes.

        Attachments

          Activity

            People

            Assignee:
            beebs Brad Bebee (Blazegraph)
            Reporter:
            bryanthompson bryanthompson
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

              Dates

              Created:
              Updated: