To connect via Zoom: Meeting ID 831-443-820
Password distributed with meeting announcement
(See instructions for setting Zoom default to join a meeting with audio and video off: https://larsoft.org/zoom-info/)
PC, Mac, Linux, iOS, Android: https://fnal.zoom.us/j/831443820
Phone:
https://fnal.zoom.us/zoomconference?m=SvP8nd8sBN4intZiUh6nLkW0-N16p5_b
H.323:
162.255.37.11 (US West)
162.255.36.11 (US East)
213.19.144.110 (EMEA)
See https://fnal.zoom.us/ for more information
At Fermilab: no in-person presence at the lab for this meeting
Release and project report
Discussion:
Tom J: can we move to using root 6.22.06b?
Has a patch that fixes a memory leak in recob::Track. Please roll this out asap.
Kyle K: available through Critic 2.06. or latest release. So comes along with art 3.08
LG: root embedded so deeply in art that we cant update root independently.
LArSoft will update along with migration to art 3.08. Schedule TBD
Herb Greenlee: Container issues and LArBatch updates
When LArSoft upgraded from art 3.05 to 3.06 (LArSoft v09_16_00), triggered lengthy discussion on larsoft listserv
Observed crashes due to missing system rpms
See Feb 2021 larsoft listserv, subject "Grid jobs and the required xxhash library"
Req'd rpms rather quickly added to std systems configs and singularity containers
Also resulted in complicated set of best practices that are required to ensure a proper singularity container is used in batch jobs
Container best practices
Showed part of the complicated, obtuse incantation required: ensures jobs land on singularity enabled nodes, and specifies the container
All of that can be contained in project.py <jobsub> xml element, so did that in this update
Disclaimer: SCD encouranging expts use the fife_utils scripts, fife_launch and fife_wrap
In mean time, still maintaining larbatch
Proposed updates to larbatch
basically, make previous best practices the default
In most cases, project.py xml element <os> will now specify a container rather than the hw OS
LArbatch PR 13
Added singularity> xml element to generate singularity condor requirement.
Defaults to "1", so enabled by default
Set to "0" if do not want that requirement
Added <cvmfs> and <stash> elements to require acces to cvmfs and cvmfs stach cache, respectively
If singularity flag is true, then interpretation of <os> xml element changes from current behavior
Will be an alias for Fermilab singularity image, or an absolute or relative path to singularity image file
No --OS option is generated in this case
Also included some pent up changes for larbatch
Command mrbslp no longer used to set up batch env if there is a local test release (always just uses plain ups setup)
Previous behavior was causing more problems than solving
Mostly won't make a difference, but see slides for a few corner cases
Copy back now done using current version of ifdhc instead of larsoft version
Needed for running some very old versions of larsoft
Changes have all been tested
Adheres to best practices using existing xml files
Works for SL6
Verified python 2/3 agnostic
PR itself is not yet approved
Discussion
ES: Asked about fife_utils fife_launch / fife_wrap
might be good to have Marc Mengel, or someone working on it to give a talk at LCM
Herb has list of use cases and questions to ask then. Forward this to whomever is speaking so that they can address them, say whether they are supported, or what is the best approach.
Noted that larbatch covers a lot of the unusual workflows that MicroBooNE uses. Hence the list above.
Will forward to me
Talk to Ken Herner regarding a speaker
LG will approve the PR