 |
 |
Attendee: Richard, John Bartelt, Wei, Booker, Chuck DQ2 and Production: all fine. Tier 2 hardware: no info, but see below AOB: Richard mentioned that nearly half of the second black box will be put in the general queues. Should ATLAS discuss with the Lab management about allocating some resources in this general pool for ATLAS users at SLAC? John Bartelt had successfully tested a reverse-proxy apache server as a front end gate of the real ATLAS web server for local replica catalogy (LRC). With this gate, we can allow outside users to access URLs reading the LRC DB, and allow inside users to access all URLs (reading and writing the DB). The gate apache is a standard, well maintained apache server. This is not in production yet but shows it is technically possible, and an workaround of a long standing security issue for WT2 operation. We are in touch with LBL and CERN to test a version of SRM v2 under FTS. Andy delivered a new component of xrootd, the cluster name service. It is an attempt to stitch together a picture of directory structure distributed on several data nodes. The first version has some issues but it looks promising. We might use it directly or build a FUSE based filesystem on top of it. In today's US ATLAS computer phone meeting, Michael Ernst expressed that he wanted all sites to have (FY08 money funded) equipment up and running by April 1. There are some confusion about a table from Jim Shank listing the pledged capacity at all sites for the next few years. Especially, are those years (2007-2011) calendar year or fiscal year? and whether those capacity excludes 20% for US physicists. Some (including SLAC) use FY08 year money to estimate 2007 year-end capacity. Other think the number on the table includes that 20% (SLAC number doesn't include that 20%). Future discussion (and new numbers from sites) may be needed to address this confusion. -- Wei Yang | [log in to unmask] | 650-926-3338(O)