| 8:30 |
Alan Silverman(Cern) |
Large System SIG
| [
Powerpoint
][
Web
]
|
| Review of recent surveys conducted by the SIG;
First impressions of the second Large Cluster Workshop;
Ideas on HEPiX-HEPNT providing technical input to grid
projects |
| 9:00 |
Ian Bird (CERN) |
LCG and HEPiX-HEPNT
| [
Powerpoint
][
Web
]
|
| How best to share information between the LCG and sites in
the HEPiX-hepnt context. |
| 9:30 |
Lisa Giacchetti (FERMI) |
AFS: What is everyone doing?
| [
Powerpoint
][
Web
]
|
| Various people have expressed an interest in learning what
other sites are doing in regard to AFS and support for it. This time slot is
reserved so that discussion can take place. |
| 10:00 |
Coffee Break |
| 10:30 |
Jean-Yves Nief (IN2P3) |
The BaBar Tier A at CC-IN2P3
| [
Powerpoint
][
Web
]
|
| A description of how the BaBar Tier A works at CC-IN2P3 (e
.g. hardware
software, problems, users' needs) as well as perspectives and the evolution
of the Tier A. |
| 11:00 |
Yannick Patois (IN2P3) |
CVS Setup at CC-IN2P3 and DataGrid 'edg-build' tool
| [
Powerpoint
][
Web
]
|
| This talk presents the way the CC-IN2P3 at Lyon setup it's CVS
servers for experiments and some custom made build tools created for the
European Datagrid Project. The first part of the talk focuses on the CVS server
setup and the second part concentrates on the edg-build tool which automatically
generates RPM packages from a remote CVS repository. |
| 11:45 |
Steve Timm (FERMI) |
Measurements of Hardware Reliability in the Fermilab Farms
| [
Powerpoint
][
Web
]
|
| [abstract] |
| 12:15 |
Lunch |
| 1:30 |
Philippe Defert (CERN)
|
Software Package Management System for the EU Datagrid
| [
Powerpoint
][
Web
]
|
| Installing and maintaining software in each node of a comp
uter
center is not a simple task. The analysis of the installation part of the f
abric
management work package will be shown which leads to new design considerati
ons.
The model developed for this work pacakge of the EU DataGrid project is the
n
presented. the design of the ASIS client updater will be changed using this
this model as a proof of concept for the future DataGrid tool suite which
could be available during 2003. |
| 2:00 |
Tim Smith (CERN) |
Managing a farm without user jobs would be eaiser
[
Powerpoint
][
Web
]
|
| A batch farm is much more than just boxes, an OS and
the raw batch scheduler code. A significant investment of effort is
required in configuring and reconfiguring to match the user
requirements and collecting, analysing and presenting user accounting
statistics. In addition diverse system protections are required to
approximate to a user sand-box environment. When the batch farms are
run as shared resources, the competition for resources, demonstration
of fairness and protection of one set of users from another complicate
the issues even further. I will chart out the course we are steering
through these problems using LSF. |
| 2:30 |
Cary Whitney (LBL) |
Linux filesystem performance via NFS
| [
Powerpoint
][
Web
]
|
| A presentation of our performance numbers for various
Linux filesystems via NFS. Our goal was to see what filesystem gave the
best NFS performance and compared that with some commercial NAS. |
| 3:00 |
Lisa Giacchetti
|
Meeting Close
| [
Powerpoint
][
Web
]
|
| 3:30 |
Coffee Break |
