Summary of December GDB, December 12th, 2018 (CERN)


Agenda

Agenda

Introduction

slides

  • 2 day ARC HTCondor at ISGC
  • Also WLCG Workshop prior to CHEP
  • SOC workshop to open registration this week
  • Also MISP workshop to joint with Zeek workshop with overlapping session

Containerised Benchmark Tools - Domenico Giordano

slides

  • Future suite should be as easy to run as HS06
  • Previous effort to use CARE to package CMS ttbar eg, tracking all files used. Manual effort.
  • Example container ~1GB in size (see inset slide 4)
  • Status of HEP workloads: good progress
  • separate runner infrastructure, since we would saturate the runner and might affect the performance for other users.

Questions

IanC: What do you think the next point to revisit?

Domenico: Progressing quite quickly, spent time to build infra, recipe properly done -> should run OK. Now second order effects, in contact with experts of experiments. Things can go quite quickly. Next to run ALICE/LHCB workloads.

IanC: Around Feb ask for another update?

Domenico: Yes

Randy: Using data to assess CPU, remove effect of storage?
Domenico: First make sure that everthing is local. Don't look only at duration but also metrics reported, CPU events, user time per events etc. Most of reports are quite detailed to decouple time spent on I/O. Don't want to just use time to run app. Case of LHCb, run full chain, final log has fine grained record of which step was run, etc. This would be part of framework - if not need to consider.

LHCONE/LHCOPN Workshop report - Edoardo Martelli

slides

  • RAL dates: TBD (IanC), work is ongoing
  • Looking glass: page to query information of different operators

Questions

Maarten: How will you do the traffic shaping to protect CMS traffic to Fermilab?

Edoardo: QoS, Juniper

Maarten: How can you recognise the right trafficf?

Edoardo: Dedicated machine

Maaretn: Have to have special machines at least at one end

Security Challenge update - Sven Gabriel

slides

  • Thanks to LHCB/Chris
  • 8-9 years ago environment was less heterogeneous
  • Responses will vary between smaller sites and larger sites with expertise/interest
  • Orchestration: Big site may take immediate action, SSC already over, so need to take care
  • Want to give site report on performance, scoring. Give best/worst/average.
  • SSC-Monitor screenshot: overview of global site status
  • Complex data model

Questions

IanC: Timeline slide, sequence?

Sven: First step is t0 - 2 weeks.

IanC: Given that you hope that the last step will take place by end of March, what's estimate for whole procedure?

Sveng: All will take 1 week, not more. Start sharp on a Monday. Not weekends/holidays. Have talks at ISGC.

Ianc: Given that, aiming for February, March?

Sveng: February.

Ianc: Interesting, seeing reponse will be interesting, testing new workflows. This work has been developing for some time now, good to see coming to fruition.

HannahS: Sites contacting VO?

Sven: No, don't really know what will happen there. Procedure is that the sites should contact us (coordinator). Vincent will not be coordinator, too much knowledge of backend. Can step in if something goes wrong. Have something similar on LHCB, be done by Chris. Sven will enjoy the show!

Asia Tier Forum report - Sang Un Ahn

slides

Data Protection discussion - David Foster, Julia Andreeva, Dave Kelsey

Report to follow

Authorisation WG pre-GDB report - Speaker: Hannah Short (CERN)

slides

Edit | Attach | Watch | Print version | History: r3 < r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r3 - 2019-01-28 - IanCollier
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LCG All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback