Polygraph 2.6.3 available

From: Alex Rousskov (rousskov@measurement-factory.com)
Date: Wed Jun 27 2001 - 15:53:20 MDT


Hi there,

    Web Polygraph 2.6.3 is available. The change log is attached.
        http://www.web-polygraph.org/downloads/

    This is a minor release that fixes a few 2.6.2 bugs and polishes
SrvLB-L7 workload. One of the fixed bugs led to floating point
exceptions at start-up time for some workloads. Please read the change
log for other details.

    Those of you using Polygraph for proxy caching tests, including
fourth cache-off preparations, please note that version 2.6.3 fixes an
important bug that may affect your hit ratios.

Thanks,

Alex.

------------- change log -----------------
2.6.2 -> 2.6.3
20010627
    - emit (errUnreachContType) "unreachable content type" when an
      embedded content type cannot be found on the [visible] server that
      hosts the container object; Polygraph can only handle embedded
      objects that reside on the same visible server as the container
    - use the same content distribution for all L7 servers (except
      for the URL extension) so that all servers will get similar
      load
    - when server's direct_access probabilities are specified manually,
      use them and do not compute our own
    - added "byte.rptm.mean", [milliseconds / kilobyte] object to Report
      Generator. The object value is visible in make_report as "byte"
      table entry in the Response Time table and as the byte_rptms plot.
    - replaced --rng_seed option with two new options: --global_rng_seed
      and --local_rnd_seed. Both default to 1. All processes within a
      test must have the same --global_rng_seed value to correctly
      generate "global" information such as a URL extension for the same
      object ID. All processes within a test should have different
      --local_rnd_seed settings to avoid lock-step behavior of things that
      should not be synchronized.
    - do not print "Potential problems" header if no problems are detected
      by make_report
    - when generating a public miss, use local public world slice
      only; using remote slices creates race condition for
      generating a miss with the corresponding remote agent; these
      race conditions lead to buggy offered hit ratio
    - W2K select(2) fails if no FDs are passed to it (i.e., all FDs sets
      are empty)
    - fixed fixed a very old bug (again) that, under certain conditions,
      left some FDs ignored by select(2); the earlier fix was ineffective
    - fixed ceil() workaround once again: the earlier fix was ineffective



This archive was generated by hypermail 2b29 : Tue Jul 10 2001 - 12:00:19 MDT