[Halld-offline] Offline Software Meeting Minutes, February 5, 2014

[Mark Ito]

Hello all,

You can find the minutes below and at

https://halldweb1.jlab.org/wiki/index.php/GlueX_Offline_Meeting,_February_5,_2014#Minutes

   -- Mark
_______________________________________________________

GlueX Offline Meeting, February 5, 2014
Minutes

    Present:
      * CMU: Will Levine, Paul Mattione, Curtis Meyer
      * FSU: Aristeidis Tsaris
      * IU: Kei Moriya
      * JLab: Mark Dalton, Gagik Gavalian, Mark Ito (chair), David
        Lawrence, Dmitry Romanov, Simon Taylor, Beni Zihlmann
      * MIT: Justin Stevens
      * NWU: Sean Dobbs

    Find a video of the meeting [32]here.

Announcements

      * We will have a data challenge meeting Friday, February 7 at 11:00
        am.

Tagged File System for Event Data

    Gagik presented his system for distribution of event data for the data
    mining project at ODU. See [33]his slides for details. He addressed:
      * Data Storage and Compression
      * Data Format
      * Data Distribution and Discovery
      * Tag-Filesystem Associated Tags
      * Tag-Filesystem Distributed System
      * Ideas (GlueX-related)

    The system is designed to retrieve a user-defined data set onto a local
    storage. There are two levels of selection: tags and filters. Tags
    identify specific sets of data files. Each file can be tagged in
    different ways and the tags may describe an arbitrary set of properties
    of the data, for example beam energy tags and and target type tags.
    Filters are routines that select individual events from each data file
    and are applied at request-fulfillment time. An example is event
    topology, e. g., the number of charged particles. The storage and
    retrieval system can be distributed over several servers and thus
    should scale.

    We ran out of time for everyone's questions to be answered. A follow-up
    session with Gagik will be arranged.

Collaboration Meeting Agenda

    Mark will ask Richard if he would like to talk about Geant4 conversion
    and related geometry changes.

Data Challenge Meeting Report

    We looked over the minutes from the [34]January 31 meeting. We deferred
    discussion of these items until the next meeting on February 7.

Reverse Magnetic Field Fix

    Simon reported on the fix he checked in. Both HDGeant and the
    reconstruction get their magnetic field via the resource mechanism of
    JANA. In particular they now use the new 1350 A field with a polarity
    reverse of what we have been using up until now. There were several
    small changes that had to made to the tracking code to accommodate this
    change.

    There is still a glitch in the acceptance at phi = 0 that seems a
    remaining corner of the backward-angle-CDC-stereo-straw problem that
    was largely fixed by Richard. Beni has confirmed the problem in a study
    where the straws were turned to lead and geantinos used to illuminate
    the CDC. Simon has sent Richard some plots documenting the problem.

Time-of-Flight Geometry Change

    Beni has updated the TOF geometry to reflect the half-width counters we
    have near the beamline. There were corresponding changes necessary in
    the simulation and reconstruction. These have all been checked in.

    There was some discussion about whether the internal numbering scheme
    for the counters was natural and consistent. No conclusions were
    arrived at about whether this even matters. Further discussion will
    follow in a smaller group.

Single-Track Test and JANA Versions

      * The single-track test failed on the most recent attempt. The test
        is now using CCDB 0.09 and JANA 0.7.1.
      * JANA 0.7.p1 has been released. It fixes a few minor issues:
           + Default directory choosing, when unspecified by the user, for
             storing resources had a bug.
           + The make-plugin script now uses scons in favor of make.

Code Validation

    Curtis asked about this issue.
      * Mark has a cron job that makes an attempt at making a performance
        benchmark on the code. Such attempts are currently failing.
      * David has worked on the suppression script for filtering sketchy
        practices flagged by valgrind.
      * This issue needs to be moved up in priority.

Action Items

      * Circulate document on database/webserver deployment developed by
        JLab IT. -> Mark
      * Research IT Division procedure for restoring a backup of the MySQL
        database. -> Mark
      * Implement pre-caching of resources. -> Mark, David
      * Ask about event format requirements for EventStore. -> Sean
      * Take the TOF geometry indexing discussion offline. -> Mark
      * Move code validation up in priority. -> Mark

    Retrieved from
"https://halldweb1.jlab.org/wiki/index.php/GlueX_Offline_Meeting,_February_5,_2014"

References

   32. 
https://halldweb1.jlab.org/talks/2014-1Q/offline_2014-02-05/lib/playback.html
   33. https://halldweb1.jlab.org/talks/2014-1Q/Hall-D-COAT.pdf
   34. 
https://halldweb1.jlab.org/wiki/index.php/GlueX_Data_Challenge_Meeting,_January_31,_2014

-- 
Mark M. Ito, Jefferson Lab, marki at jlab.org, (757)269-5295