DataPlacementMinutesMar21

Introduction

• Friday, March 21, 1pm EST
• (605) 475-6000, Access code: 735188
• Attending: Michael, Kaushik, Hiro, Wensheng, Rob

Discussion

• Discussion points, see: Presentation1.ppt
• Try to improve the replication exercises T1-->T2
• Take scripts from Alexei and create new monitoring pages by release number.
• We control replication with the BNL cloud. Process will be: Kaushik will provide scientific oversight of which datasets need replicating. Send these high level requests to the team of Hiro and Wensheng who will manage the replication within the cloud.
• Hiro is already creating a website that focuses on T1-->T2, but including troubleshooting detail, and automatically re-scanning. Hope to have prototype by Wednesday's meeting. Kaushik will send an initial list of datasets for replication.
• One issue is how to aggregate views - how to handle very large datasets.
• Will revisit the issue of dataset deletion. User datasets will be implemented so that they can be managed from pathena.
• Discussed a fully automated system for managing holes - for the development project. Not practical for site admins to track down missing files by visually scanning webpages. Need for automation.
• As part of the interim project, Hiro will send test subscriptions to each site, and will provide a manual deletion script that sites can cron.
• Agreed we would view how the system evolves and revisit if necessary for features/changes.

-- RobertGardner - 19 Mar 2008