Analytics WG meeting

Europe/Zurich
31/S-023 (CERN)

31/S-023

CERN

22
Show room on map

Participants: Dirk Duellmann, Christian Nieke (minutes), Luca Menichetti, Rainer Toebicke,
Sebastian Ponce, Luca Canali, Pedro Andrade, Raul Gorcia Martinez, Ulrich Schwickerath, Bernd Panzer-Steindel,
Domenico Giordano, Luca Magnoni

News:
- TWiki created by Christian to collect information on data sets at https://twiki.cern.ch/twiki/bin/viewauth/ITAnalyticsWorkingGroup/WebHome
- Experiments asked for cooperation and access to analysis
infrastructure and data
- Putting non-CERN data from WLCG there will have performance/storage impact
- will setup meetings with experiment analytics representatives to discuss details

LAN DB - What metrics may be useful for analysis?:
- Talk by Raul Garcia Martinez
- Network Services will move extracts of their data to Hadoop in some
months.
- Focus is on Topology, later error reports and quantitive values (like
throughput)
- Confidentiality might be an issue, but we might get a subset of e.g.
only machines in the computing centre


Experiment dashboards - existing data, analysis and plans:
- Talk by Luca Magnoni
- SDC will move from the DB hadoop cluster to the analystics hadoop cluster
- Start with xroot transfer data (main scalability proof)
- Should we have a common way to store data?
   - Yes, but later, after we tested the implications of different formats
   - For now, we stick with any kind of raw format, which could be
transformed
- Should we use the same technology to collect data ? (e.g. flume, stomp ..)
   - This would make it easier to adapt into different formats later
   - This will be discussed in smaller groups
- Ulrich will arrange for putting LSF accounting/config data to analytics cluster

There are minutes attached to this event. Show them.
    • 14:00 14:05
      Minutes & News 5m
      Speaker: Dirk Duellmann (CERN)
    • 14:05 14:25
      LAN DB - What metrics may be useful for analysis? 20m
      Speaker: Raul Garcia Martinez (CERN)
      Slides
    • 14:25 14:45
      Experiment dashboards - existing data, analysis and plans 20m
      Speaker: Luca Magnoni (CERN)
      Slides
    • 14:45 15:05
      Data import 20m