GlueX Software Meeting, November 26, 2019

From GlueXWiki
Revision as of 18:29, 29 November 2019 by Marki (Talk | contribs) (add minutes)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

GlueX Software Meeting
Tuesday, November 26, 2019
2:00 pm EST
JLab: CEBAF Center A110
BlueJeans: 968 592 007

Agenda

  1. Announcements
    1. New version set: version_4.11.0.xml
    2. "atime" enabled on the work disk (/work/halld)
    3. Work disk still full
  2. Review of minutes from the last Software Meeting (all)
  3. Review of minutes from the last HDGeant4 Meeting (all)
  4. Accidental scaling factors in DSelectors (Sean)
  5. Review of recent issues and pull requests:
    1. halld_recon
    2. halld_sim
    3. CCDB
    4. RCDB
  6. Review of recent discussion on the GlueX Software Help List (all)
  7. Action Item Review (all)

Slides

Talks can be deposited in the directory /group/halld/www/halldweb/html/talks/2019 on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2019/ .

Minutes

Present:

  • CMU: Naomi Jarvis
  • JLab: Thomas Britton, Mark Ito (chair), Igal Jaegle, David Lawrence, Keigo Mizutani, Justin Stevens, Simon Taylor, Beni Zihlmann

There is a recording of his meeting on the BlueJeans site. Use your JLab credentials to access it.

Announcements

  1. New version set: version_4.11.0.xml The latest set of default version was released on November 19.
  2. "atime" enabled on the work disk (/work/halld). Until yesterday, files on the work disk have not had their access times updated by the filesystem in order to boost performance. Yesterday atime updates were turned on on /work/halld. to be precise "relatime," a ZFS feature was enabled. Atimes less than 24 hours in the past are left alone. Having this feature will help us distinguish between old files that have been used recently and those that have gone unread for a long time.
  3. Work disk still full. The work disk is at 98% full see the plot below for recent history and follow this link to see a report on work disk usage.
Work Disk Usage
  1. Mirrors of CCDB SQLite Files. Thomas pointed out that there are still some MCWrapper users out there using old versions that are pointing their JLab Farm jobs to the SQLite files on the work disk. The work disk has a known pathology with file locking that causes these jobs to hang. He asked that those files be deleted.
    • David suggested that we remove some legacy magnetic field maps from the CCDB. They were kept in to allow old code to access them there, but that code is now many years old. Field maps are now accessed as a JANA Resource. Deleting them will save a lot of space in the database, and in particular in the SQLite files generated from MySQL, perhaps a factor of two or more. Mark will look into doing this.
    • Added in press: the deed was done. See this email message from Mark for more details.

Review of Minutes from the Meeting on October 29

We went over the minutes.

Beni emphasized the issue of upgrading from Python 2 (P2) to Python 3 (P3). [Added in press: he refers us to this link on the subject].

  • We run with a mixture of the two, scattered all over the source code base.
  • David pointed out last time that conversion from P2 to P3 mostly consists of changing "print" statements.
  • He also told us that Python 2 official support ends in January 2020.
  • He has been using #!/usr/bin/python3 to identify scripts that are meant for the P3 interpreter.
  • And SCons uses P2 at present.
  • P2 vs. P3 differs from C++ vs. C++11 in that the compiler choice is made before building the entire code base and so builds fail with the wring choice. Whereas the interpreter choice is important only at run-time allowing for more flexibility. In addition, the choice of interpreter can be made explicitly on the command line, e.g., /path/to/python script.py.

The clear path forward is still not visible.

Review of the HDGeant4 Meeting on November 19

We went over the minutes. There were no significant comments beyond those already in the minutes.

Accidental scaling factors in DSelectors

The scaling factors, produced by Beni and Richard Jones, that correct for less-than-100% duty factor of the machine are now available in the CCDB thanks to Sean Dobbs.

New CentOS 7.7 Nodes on the JLab Farm

Beni brought up the subject of using the new nodes on the farm, announced on November 15 running with CentOS 7.7. He was not able to get a successful build on ifarm1802 (recently upgraded to 7.7); there was a problem with the MySQL libraries. Justin pointed out that if we get a public build going on these nodes we can use them for monitoring launches to avoid competition with other tasks. Mark will look into a build on this platform.

New Translation Tables for Fall 2019

Justin reported that new translations tables are on the way. They incorporate the upgraded TOF, the GEM/TRD channels, and a few new DIRC-related channels.

Using Google Groups for Email Forums

We briefly discussed some of the problem with using Google Groups for email discussions. There are many positive aspects which motivated the switch in the first place, but down-sides have been noted:

  • Use of non-google emails is a bit complicated and has limitations (in particular where authentication is needed).
  • The group use of a single account to own the groups, accessible to the entire Collaboration is running into occasional security measures from Google.
  • Authenticated web access from some universities (including CMU) using Google Suite is an on-going problem.

CLAS12 and the Geant4 Collaborations are moving to a product called Discourse. Trello was mentioned as another options. We probably would have to lobby, in concert with other groups at the lab, in order to get IT Division to pay for a site license for any of these products. This is not a crisis at this point, but it is something for us to think about.