Virgo Kick-Off Meeting on Computing

Europe/Madrid
Parc Científic de Barcelona
Description

Accommodation Information:

A list of nearby accommodation locations can be found here.

Please book your stay as soon as possible. The week after is the Mobile World Congress and Barcelona gets really crowded.

Venue Information:

The venue is located here

Access map:

    PCB Access Map

To enter the building you must request a visitor badge in the reception desk (i) indicating our institute (ICCUB) and any of us as contact person: A.Casajus, J.Castañeda, J.Portell or J.Trenado.

The meeting will take place in the 3rd floor of "Torre D" (tower on the right at the end of the main corridor where the reception is located).

Remote connection:

Topic: PCB
Time: Feb 21, 2019 9:00 AM Madrid
 
Join Zoom Meeting
 
One tap mobile
+16699006833,,753093200# US (San Jose)
+19292056099,,753093200# US
 
Dial by your location
        +1 669 900 6833 US (San Jose)
        +1 929 205 6099 US
Meeting ID: 753 093 200
Find your local number: https://zoom.us/u/adL7Tsg29q
 
Registration
Participants
Registration
Participants
  • Adrià Casajús
  • Andres Tanasijczuk
  • Franco Carbognani
  • Gabriele Gaetano Fronzé
  • Javier Castañeda Pons
  • Jordi Portell de Mora
  • Juan Trenado
  • Luca Rei
  • Pablo Barneo
  • Sarah Caudill
  • Stefano Bagnasco
There are minutes attached to this event. Show them.
    • 09:30 10:00
      LVC collaboration 30m

      How do Virgo and Ligo share/compete for resources
      - Who can take decissions on sw, data management, sites...
      - For example, should virgo use ligo's git?
      - Offline sw is shared amongst both collaborations?
      - How is the analysis sw developed? Who 'owns' the sw?

    • 10:00 11:00
      SW management 1h

      Current sw management rules at: https://wiki.virgo-gw.eu/Software_Management/WebHome

      Which sw is managed?
      - Online?
      - Framework?
      - User analisis?

      Role of the existing Ligo-Virgo Software Configuration Control Board (SCCB) in this game: https://wiki.ligo.org/Computing/SCCB

      /virgoApps more than 900 packages!
      - Review product/module breakdown

      Is there a "Testing environment"?, useful/needed to decouple from the operational one.
      - Not only for data analysis pipelines, but also at ‘framework’ level; e.g. to test Rucio, DIRAC, data exchange, bookkeeping, workload…

      How are the main pipelines developed? Internal group effort or collab effort?
      - Which are developed by Virgo?
      - Which are coming external to Virgo?

      SW distribution
      - Conda (Review of the corresponding proposal https://tds.virgo-gw.eu/ql/?c=13967)
      - Cvmfs
      - Singularity

      Newer toolings available
      - Git
      - If/when/how
      - CMT not maintained
      - Can we take the conda migration as an opportunity to replace cmt with something else? CMake, bazel…

      Automated release, packaging and distribution possible?
      - Roadmap

    • 11:00 11:15
      Coffee break 15m
    • 11:15 13:00
      Offline activities 1h 45m

      Two main topics:
      - Data management
      - Workload management

      Tools to fit Virgo needs (Rucio/DIRAC/others)
      LIGO goes to Rucio - thus Virgo should, at least, be able to handle it.
      Pros-cons of tools to have an overall picture and ideally take a decision and run with it.

      Running LIGO-Virgo analyses on Virgo computing resources: status & plans

      Data storage
      - Which data is stored where? Virgo raw only in cascina? h(t) everywhere?

      Rewiew of the current relationship with Computing Centers.

    • 13:00 14:30
      Lunch 1h 30m
    • 14:30 16:00
      Online processing 1h 30m

      Clarifications on data flow diagrams and processing modules
      - Computing just on raw Virgo data, or also on raw LIGO data?
      - Just on h(t), both LIGO+Virgo?.
      - DetChar/Quality: just on raw Virgo? Also on LIGO?
      - Processing raw data generates h(t), and also veto?
      - General review of data flow diagram to identify possible problems or limitations

      Migration/upgrade processes ongoing: storage, Kafka, etc.
      - Deadlines, tracking system, etc
      - For the Open Public Alerts (OPA) machinery (gracedb, gwcelery, lvalert, …) we are fully relying on Ligo, are we happy about it?
      - Bulk data handling (envision the future Data Transfer and Access integrated solution based on a EGI/OSG Framework)
      - GraceDB is managed by Ligo, are we happy with that?

    • 16:00 16:15
      Coffee break 15m
    • 16:15 17:30
      Wrap-up and actions to be taken 1h 15m