Agenda
Spring 2004 HEPiX meeting at the eScience Institute, Edinburgh
24-28 May, 2004

Agenda last modified 09-May-2007 13:36

The HEPiX home page may be found here

The meeting home page and registration may be found here

A list of registered participants may be found here

note: Links to Video recordings called "streaming" are hosted on a server capable of multi-speed video streaming.
Links called "download" are on a non-streaming server

Monday 24 May - HEPiX - Day 1
08:30-09:30   Arrival and Registration

Video

09:30-09:40   David Kelsey (RAL) / Welcome and Introduction  No video
09:40-10:00   Dave Berry (NeSC) / Introduction to NeSC  No video
10:00-10:15   Tony Chan / BNL Site Report  download
 streaming
10:15-10:30   Juergen Baschnagel / PSI Site Report  download
 streaming
10:30-11:00 Coffee  
11:00-11:15   Chuck Boeheim / SLAC Site Report  download
 streaming
11:15-11:30   Stephan Wiesand / DESY Site Report  download
 streaming
11:30-11:45   Michel Jouvin / LAL Site Report  download
 streaming
11:45-12:00   Mark Kaletka / FNAL Site Report  download
 streaming
12:00-12:15   Kelvin Edwards / JLAB Site Report  download
 streaming
12:15-12:30   Andrei Maslennikov / CASPUR Site Report  download
 streaming
12:30-14:00 Lunch  
14:00-14:15   Cary Whitney / NERSC Site Report  download
 streaming
14:15-14:30   Pierre-Francois Honore / DAPNIA Site Report  download
 streaming
14:30-15:00   Helge Meinhard / CERN Site Report download
 streaming
15:00-15:15   Helmut Kreiser / GSI Site Report download
 streaming
15:15-15:30   Corrie Kost / TRIUMF Site Report download
 streaming
15:30-16:00 Coffee  
16:00-16:10   Sabah Salih / Manchester Site Report  download
 streaming
16:10-16:20   Peter Gronbech / Oxford Site Report  download
 streaming
16:20-16:35   Andrew Sansum / RAL Site Report  download
 streaming
16:35-18:00   Nathan Jones (Red Hat) / The Migration to Red Hat Enterprise Linux

Nathan Jones, one of the US Sales Directors, will discuss migrations from Red Hat Linux in the HEP community and the current options for migrating to Red Hat Enterprise Linux. Nathan will discuss some details on those Labs that he knows are currently migrating. He will discuss all of the current solutions available from Red Hat (including Fedora). He will discuss Red Hat's commitment to the HPC and HEP communities; Red Hat's development of a pricing structure that understands this market; future product strategies; and company direction. He will also be willing to take comments and suggestions back on what the HEP community needs from Red Hat (i.e. in terms of new partnerships, functionality, or business terms) as it grows it's compute requirements.

 No video
Tuesday 25 May - HEPiX - Day 2
09:00-09:30  German Cancio (CERN) / ELFms deployment in- and outside CERN CC

ELFms (http://cern.ch/elfms) stands for Extremely Large Fabric management system, and comprises tools for fabric configuration and installation (Quattor), monitoring (Lemon), and hardware/state management (LEAF). In this talk, I will describe the ELFms modules and overall architecture, as well as the integration/deployment status at CERN-CC and other sites/projects.

 download
 streaming
09:30-09:50  Miroslav Siket (CERN) / Web presentation of Lemon monitoring information

On top of the Lemon infrastructure to gather status and performance information on the computers within a computer center, we have implemented a framework to make those data easily accessible to the users through the web based interface. The solution is based on the RRD tools for storing and PHP/Apache for presentation purposes. Overview of the current status and the undergoing activities will be given.

 download
 streaming
09:50-10:15   Paul Millar (Glasgow) / Lessons Learnt in the Management of CDF Hardware and Service Contract
  (Authors: Morag Burgon-Lyon, A. Paul Millar, Richard St.Denis, A. Stan Thompson)

The CDF Experiment has four identical clusters deployed at different UK institutions. A RAID failure on the Glasgow University cluster led to over 2 months of downtime. We discuss the problems encountered and actions taken by both ourselves and the vendor. We present a improved procedure for tackling future hardware problems.

 download
 streaming
10:15-10:40   Maria Dimou (CERN) / LCG User Registration and VO management

The talk will explain the status of the requirements investigation for LCG User Registration and VO management as mandated by the LCG Grid Deployment Board. The registration procedures used at present will also be described.

 download
 streaming
10:40-11:10 Coffee  
11:10-11:30   Gilbert Grosdidier (LAL) / LCG Testing Suites (full talk)
 presented by Michel Jouvin (LAL) - the slides as presented

The LCG/EDG Testing Suites will be described, together with their framework. The aim will be to show how powerful they can be to spot failures on a running testbed, and to to push sysadmins to use them on their own site, to detect install oddities and possible misconfigurations.

 download
 streaming
11:30-12:00   Oliver Keeble (CERN) / LCG Update

LCG-2 has been deployed successfully, integrating over 50 sites and giving access to more than 3000 CPUs. The LHC experiments' data challenges have put the production environment under heavy load and revealed a number of operational issues requiring attention. This talk reports on the modifications made or scheduled in response to the data challenges, presents a summary of current LCG-2 status and describes future developments.

 download
 streaming
12:00-12:30   Tony Doyle (Glasgow) / GridPP Project Status and Outlook

GridPP is a collaboration of Particle Physicists and Computing Scientists from the UK and CERN, who are building a Grid for Particle Physics. GridPP involves 19 UK Universities plus CCLRC and CERN and is embedded in a complex environment involving many other large organisations such as the CERN LHC Computing Grid (LCG) Project, the European DataGrid/EGEE, US Grid Projects and the large international particle physics experiments. The talk will cover the current deployment status, middleware developments in the UK, and the plans for GridPP phase 2.

 download
 streaming
12:30-13:45 Lunch  
13:45-14:10   Connie Sieh (FNAL) / SLiding into Free Enterprise Linux

There is a need for a common, freely available, enterprise quality Linux distribution for use among the High Energy Physics community. The current enterprise Linux distributions cost too much for many sites to afford. Linux is not always compatible between different Linux distributions. This makes it difficult for one scientist to share their code with another at a different site. The system administration also becomes easier when a common distribution is used.

We have created "Scientific Linux" (SL) to fulfill this need. It is a community driven, freely available enterprise Linux. It is useful straight out of the box, and it is also easily customizable for the various High Energy Physics sites. Scientists are able to customize "Scientific Linux" for their own site, while maintaining compatibility with other sites. This will allow the various sites to easily share their software via a central repository. "Scientific Linux" also includes enhancements that are of value to the High Energy Physics community that are not available in commercial enterprise Linux distributions.

 download
 streaming
14:10-14:30  Connie Sieh (FNAL) / Using YUM to increase security of Linux  download
 streaming
14:30-15:00  Jaroslaw Polok (CERN) / Next CERN Linux distribution

Overview and details of the next CERN Linux distribution.

 download
 streaming
15:00-15:30  Steven Timm (FNAL) / Infrastructure and provisioning at the Fermi High Density Computing

Construction has begun to prepare the Fermi High Density Computing Facility. This facility will eventually host more than 3000 computers over the next five years. This presentation will discuss the infrastructure plans for this facility, and the plans Fermilab system administrators are making for provisioning and managing the systems as they come in to this remote location.

 download
 streaming
15:30-16:00 Coffee  
16:00-16:20   Manuel Guijarro (CERN) / Report on CVS services

In production since more than 2 years, IT/PS central CVS services host over 90 software projects of various sizes. This presentation presents the 2 architectural approaches used at CERN, as well as data about service status and ongoing developments.

 download
 streaming
16:20-16:50   Stephan Wiesand (DESY) / The new way of handling Unix application software at DESY

We started providing Software for our strategic platforms Linux and Solaris in the form of RPM packages. Packages are built according to a set of rules and a new filesystem layout to allow installation handling by automated tools. Installation can be fully local or on a central filesystem (AFS) or anything in between, configurable per system, package and version.

 download
 streaming
16:50-17:20   Brian Scott (SLAC) / Citrix and remote access at SLAC  download
 streaming
17:20-18:00   Bob Cowles (SLAC) / Recent Security Threats and Vulnerabilities  download
 streaming
Wednesday 26 May - HEPiX - Day 3
09:00-09:20   Yves Fouilhé (CC-IN2P3) / BQS Update

BQS is the batch system that we are using at the Centre de Calcul de l'IN2P3. I will present the most recent evolutions of BQS, eg. support for parallel and arborescent jobs, and jobs submitted from grids, including LCG2.

 download
 streaming
09:20-10:05   Cary Whitney (NERSC) / New Developments on PDSF

I would like to talk about the latest developments on PDSF and their effects on the users. This ranges from reclaiming old hardware to security and software modifications to allow a wider range of jobs and environments to function together. (CHOS, ProcDN, St Michael/Patchfinder, Kernel hotfixes, One-wire, Capacitors, Event monitoring, SGE/LSF, future projects.)

 download
 streaming
10:05-10:30   Ignacio Reguero (CERN), presented by Manuel Guijarro (CERN) / Overview of Solaris Issues at CERN

My presentation will review the main questions around the Solaris platform at CERN and the activities of the Solaris Support Service. The items that will be considered are
- Solaris 9 Certification
- Compiler Status
- Quattor (EDG WP4) deployment
- Blade server tests and N1 Management
- EDG WP4 Monitoring on Solaris
- Long term plans for Solaris Support

 download
 streaming
10:30-11:00 Coffee  
11:00-12:30   Red Hat Linux Panel and Discussion (convened by Alan Silverman)

This session discusses how the major labs will support Redhat Linux in the future. There will be short presentations from SLAC, Fermilab and CERN followed by an open discussion.
Jan Iven (CERN)
Mark Kaletka (FNAL)
Chuck Boeheim (SLAC)

 No video
12:30-13:45 Lunch  
13:45-14:00   Michel Jouvin (LAL) / Anti-SPAM experience at LAL

This talk will present the integration of SpamAssasin and Sendmail at LAL and the tuning done to achieve a high rate of SPAM detection without false positive. In particular, the experience with bayesian analysis will be presented.

 download
 streaming
14:00-14:30   Wolfgang Friebel (DESY)/ AFS file space administration with ARC version 2

A new version of "Authenticated remote control" (originally by R. Toebbicke,CERN) has been developed. It uses SASL for authentication and supports GSSAPI (Kerberos5, PKI), Kerberos4 and other authentication mechanisms. The AFS file space administration program used at CERN (afs_admin) was designed for the old arc and did support Kerberos 4 only. It has been partly rewritten to work now with ARC v2. Other arc applications like batch token prolongation have been ported as well.

 No video
14:30-15:00   Alf Wachsmann (SLAC) / Report from the AFS Best Practices Workshop

I will give a summarizing report from the first AFS Best Practices Workshop held at SLAC.

 download
 streaming
15:00-15:15  Michael Draper (CERN) / InDiCo - An Integrated Conference Management Tool  download
 streaming
15:15-15:30  Michel Jouvin (LAL) / SUS integration with SMS

This talk will present our experience with Update Feature Pack for SMS which provides SUS/WindowsUpdate functionality through SMS. 

 download
 streaming
15:30-16:00 Coffee  
16:00-16:30   Tim Smith (CERN) / Developments in the CERN Disk and Tape Storage Services

Early in 2004 the tools and techniques used to manage disk, tape, and stage servers were refreshed in adopting the QUATTOR tool set. The LEMON monitoring suite was deployed to raise alarms and gather performance metrics. Monitoring has been extended from the OS and application level down into the HW and environment. These developments will be presented in addition to operational experiences such as the widespread replacement of disks to regain managability of the disk servers and the large scale tape media migration to permit old tape drive technologies to be retired.

 download
 streaming
16:30-17:10   Jan Iven (CERN) / Disk Storage performance and high-speed interconnects

Applied research in CERN-IT-ADC * Diskserver perfomance evolution (on behalf of P.Kelemen) * Fast local interconnects (on behalf of A.Horvath) * USB/Firewire-attached storage (on behalf of A.Hirstius)

 download
 streaming
17:10-17:30   Wojciech Wojcik (IN2P3) / Conclusion of the HEPiX days and plans for future meetings download
 streaming
19:30 Conference Dinner - Howies Waterloo Place, 29 Waterloo Place, Edinburgh  
Thursday 27 May - Mass Storage Workshop - Day 1
08:58   David Kelsey (RAL) / Welcome and Logistics  No video
09:00-09:50   Paul Bradshaw (IBM) / IBM SAN File System for GRID and HPC Environments

This talk will provide an overview of the recently announced IBM TotalStorage SAN File System (Storage Tank) and how it is being used in the GRID and HPC environments to meet their needs. Discussions on how the system is being used at CERN and other locations will be highlighted, as well as some future directions.

 download
 streaming
09:50-10:40   Gordon Kennedy (STK) / Information Life-cycle Management

The application of an information life-cycle management strategy to large-scale storage requirements

 download
 streaming
10:40-11:10 Coffee  
11:10-11:50   Andrei Maslennikov (CASPUR) / New results from CASPUR Storage Lab

We will be reporting several results obtained in the last 6 months. This time we have tried disk systems from Datadirect, IBM and Infortrend and new fibre channel equipment from Qlogic, along with the different file system solutions. The work was done in collaboration with CERN and ENEA.

 download
 streaming
11:50-12:20   Wayne Schroeder (UCSD) / Overview of the SDSC Storage Resource Broker  download
 streaming
12:20-13:30 Lunch  
13:30-14:15   Jens Jensen (RAL) / Experience of supporting multiple interfaces (SRM/SRB/SE)  download
 streaming
14:15-15:00   Jean-Philippe Baud (CERN) / GFAL and LCG data management  download
 streaming
15:00-15:30 Coffee  
15:30-17:30   Mass Storage & WAN Data Movement Forum discussion
  “LHC (raw/ESD) data broadcasting CERN à Tier-1”
  LCG Service Challenges Slides from Ian Bird (CERN)
  WAN RAW/ESD Data Distribution for LHC Bernd Panzer-Steindel (CERN)
  • Requirements
  • Tier-1 responsibilities
  • Networking issues
  • Architecture
  • Timeline

 (Link to document re LCG MOU Task Force. Tier1 services and responsibilities)
 

 download
 streaming
Friday 28 May - Mass Storage Workshop - Day 2
09:00-10:00   Mass Storage & WAN Data Movement Forum discussion
  “LHC (raw/ESD) data broadcasting CERN
à Tier-1”
  • Wrap-up from day 1
  • Next steps towards a real test-setup

  Les Robertson (CERN) / Data Management Service Challenge
  Olof Barring (CERN) / Exporting Raw/ESD data from Tier-0 -> Tier-1s
 

 No video
10:00-10:25   Jon Bakken (FNAL) / Storage system integration with high performance networks  download
 streaming
10:25-10:50  Doris Ressmann (FZK) / Integrating dCache into an existing tape storage environment at GridKA.

The dCache system developed at Desy and FNAL is recently introduced to LCG for the connection of online to background storage. At GridKa dCache was installed to provide an initial 2 PB tape storage in 2007, which will grow exponentially afterwards. The existing IBM Tivoli Storage Manager environment and LTO libraries are used through a dedicated dCache to tsm driver. Presented are the initial results and limitations of the new dCache/TSM connection. 

 download
 streaming
10:50-11:15 Coffee  
11:15-12:00   Olof Bärring (CERN) / CASTOR SRM v1.1 experience  download
 streaming
12:00-12:30   Alan Silverman (CERN) / Mass Storage Workshop summary  download
 streaming
12:30-13:30 Lunch