[Halld-offline] Software Meeting Minutes, November 24, 2020

[Mark Ito]

Folks,

Please find the minutes here 
<https://halldweb.jlab.org/wiki/index.php/GlueX_Software_Meeting,_November_24,_2020#Minutes> 
and below.

   -- Mark

________________________________________


    GlueX Software Meeting, November 24, 2020, Minutes

Present: Alex Austregesilo, Thomas Britton, Sergey Furletov, Mark Ito 
(chair), Igal Jaegle, Naomi Jarvis, Simon Taylor, Nilanga 
Wickramaarachchi, Beni Zihlmann

There is a recording of his meeting 
<https://bluejeans.com/s/KwdHeiR6vUQ/> on the BlueJeans site. Use your 
JLab credentials to authenticate.


      Announcements

 1. Build Scripts Releases 2.4
    <https://github.com/JeffersonLab/build_scripts/releases/tag/2.4> and
    2.5
    <https://github.com/JeffersonLab/build_scripts/releases/tag/2.5>.
    See the release notes (linked) for the changes. Beni reported a
    successful built on Ubuntu 20.
 2. New version set: version_4.32.0.xml
    <https://mailman.jlab.org/pipermail/halld-offline/2020-November/008385.html>.
    Released last night. First version set with changes for Python 2/3
    compatibility described in Mark's talk below.


      Review of Minutes from the Last Software Meeting

We went over the minutes from the meeting on November 10th 
<https://halldweb.jlab.org/wiki/index.php/GlueX_Software_Meeting,_November_10,_2020#Minutes>. 


  * The work disk is still very full. Mark is working on producing
    personalized reports for power users to identify possible areas for
    archival and/or deletion.
  * Thomas reported that pilot jobs from the OSG have landed at JLab,
    but they don't run Edgar Fajardo and Wesley Moore are working on it.
  * Beni reported that halld_recon issue #462
    <https://github.com/JeffersonLab/halld_recon/issues/462> *TOF dE/dx
    written as all zeros in monitoring hists and ROOT trees since
    2020-07-10* has been resolved.
  * Nilanga reported that halld_recon issue #464
    <https://github.com/JeffersonLab/halld_recon/issues/464> *Issue with
    code for chi square comparison of two trees* has been resolved.
  * The Software Working Group sends its congratulations to Dmitry and
    Olga on the birth of their son.


      Moving Forward to New Linux Distributions

Mark described work in two areas, Python 2/3 compatibility and 
Singularity container construction, that will enable us to upgrade 
supporting software packages (e.g., Geant4) and distribute a uniform 
Linux environment as well as exploit new features of C++ and Python. For 
details seehis slides 
<https://halldweb.jlab.org/doc-public/DocDB/ShowDocument?docid=4793>.

  * Python 2/3 compatibility: a new Python module, python-future, is
    used to allow single Python files to be used under both Python 2 and
    Python 3.
  * Singularity container construction: a new Git repository,
    hd_singularity <https://github.com/JeffersonLab/hd_singularity>, is
    available with recipes and a script to build containers that will
    run GlueX software.


      Review of recent issues and pull requests

We went over the flurry of MCwrapper issues 
<https://github.com/JeffersonLab/gluex_MCwrapper/issues?q=is%3Aissue> 
submitted by Naomi last week. Most of the time was spent on Issue #46 
<https://github.com/JeffersonLab/gluex_MCwrapper/issues/46> *SubmitSim 
needs info on how to supply command line arguments for the generators*. 
The discussion boiled down to where to draw the line between info that 
MCwrapper should provide and documentation generated by the authors of 
event generators.

On a side note, Naomi has pioneered the technique of writing a 
subdirectory specific README.md 
<https://github.com/JeffersonLab/halld_recon/blob/master/src/plugins/monitoring/CDC_dedx/README.md> 
that could be used for documenting generators.


      Stability of DSelector on batch farms

Alex posted notes on a solution to a problem 
<https://mailman.jlab.org/pipermail/halld-offline/2020-November/008386.html> 
that arises when running multiple proof sessions, for example in a set 
of batch farm jobs. Not all threads complete, or entire jobs may quit.


      Action Items

 1. Ask the Computer Center about staging the start of jobs in time. (Mark)
 2. Make sure that the automatic tests of HDGeant4 pull requests have
    been fully implemented. (Mark, Sean)
 3. Finish conversion of halld_recon to use JANA2. (Nathan)
 4. Ping Dmitry on CCDB Issue #68. (Mark)
 5. Remind folks about their problematic cache files. (Alex)

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halld-offline/attachments/20201124/2db1b6f8/attachment.html>