AtlasFdr1

From GridPP Wiki
Jump to: navigation, search

Purpose

This page gathers information which is used to show where data from the ATLAS Full Dress Rehearsal will be placed within Tier-2 sites in the UK.

DPDs

Personally produced DPDs should be registered at AtlasFdr1Dpd

FDR-1 placement at Tier-2s

Decision has been made 13/2/08 based on information received in time

  • Placement is based on a combination of the physics requests made by the groups by 13/2/08 (in the table at the bottom of the page) and the readiness state of the sites as determined by ATLAS UK computing. Please see notes below for more information.
  • This is for FDR-1 only. We will do a further iteration for FDR-2.
Region Site Stream
London ICL EGamma
North LANCS Muons
North LIV EGamma
North MANC Jets
North SHEF MinBias
Scot UKI-SCOTGRID-GLASGOW_DATADISK MinBias
South UKI-SOUTHGRID-BHAM_DATADISK Jets
South CAM Muons
South UKI-SOUTHGRID-OX-HEP_DATADISK Jets
South RALPP EGamma

Responsible - Alan Barr, Roger Jones, Tony Doyle

Trigger to Stream matching for Full Dress Rehearsal-I

For information on the streams see: http://tbold.web.cern.ch/tbold/view_menu.php?name=lumi1E31_13.0.40

For information on the FDR in general see: https://twiki.cern.ch/twiki/bin/view/Atlas/FullDressRehearsal

Information Gathering from Sites for FDR-I:

FROZEN 13/2/2008 - AJB

The streams are: EGamma, JetTauEtmiss, Muons, MinBias (+ not for physics: Calibration, Express)

The notes for the sites were as follows

  • This placement is for FDR-I only, and will change for FDR-II
  • The purpose is to do as complete a test as possible of the distributed computing model, so your data of interest may not be (indeed is likely not to be) at your home site. This is deliberate! All the data would fit easily at every site, but if we put it there we do not test the tools we will need for larger data-sets.
  • Each Tier-2 cloud will hold a complete set of AOD, but, in order to test the distributed computing model, not all sites will do so.
  • To reduce traffic, we will try to make some mapping between physics interets and data placement
  • List your site's physics interests in some sort of order of priority, depending on likely maximal usage from people at your site.
  • The initial list is based on sketchy partial initial information. If it is wrong, please dont complain, but fix the table for your institute and send me a message to say what you have done.
  • Please include between one and three streams. The actual number of streams placed will depend on the relative data volumes and other factors.
  • If possible, include approximate number of people analysing the data (of this FDR) in brackets after stream name.
  • FTE is the number of people at your site who will be actively involved in analysing the data for this FDR on that' stream. If people are working on more than one stream, divide their FTE across streams.
  • SRM2.2 is the Storage Resource Manager version required for participation in the FDR. Your site needs to have this installed.


Sites were requested to list their preference for which streams to host in order follows.


Tier-2 Site SRM2.2? Stream1 FTE Stream2 FTE (Stream3) FTE
London QMUL No Muon - EGamma - - -
London RHUL No EGamma - - - - -
London UCL No Jets 3 eGamma 2 Muons 1
London IC-LeSC Yes Jets 3 eGamma 2 Muons 1
London Brunel-44 Yes eGamma 2 Muons 2 Jets
North Lancaster Yes Muons 1 Min Bias 0.5 - -
North Liverpool  ? EGamma - - - - -
North Manchester  ? - - - - - -
North Sheffield  ? Jets - MinBias - EGamma -
Scot Glasgow Yes * MinBias 3 Muon 2 EGamma 2
South Birmingham Yes EGamma 1 Jet 0.5 Muon 0.5
South Cambridge No - - - - - -
South Oxford Yes EGamma 3 Jets 2 - -
South RAL PP Yes EGamma 2 - - - -

Is your site in a suitable state to receive data?

Placement of data will be prioritised at sites which are functional. To check your site see:

More information about service tests:


Responsible: Alan Barr