From - Fri Apr 5 13:50:02 2002 Return-Path: Received: from smtp2.libero.it (smtp2.libero.it [193.70.192.52]) by castor.ts.infn.it (8.11.6/8.11.6) with ESMTP id g3574aY08200 for ; Fri, 5 Apr 2002 09:04:36 +0200 (MET DST) Received: from ts.infn.it (151.25.60.117) by smtp2.libero.it (6.5.015) (authenticated as stefano.belforte@iol.it) id 3C99A71000A5415C; Fri, 5 Apr 2002 09:03:17 +0200 Message-ID: <3CACCA50.C2F37F2E@ts.infn.it> Date: Thu, 04 Apr 2002 23:49:04 +0200 From: Stefano Belforte X-Mailer: Mozilla 4.78 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: r.stdenis@physics.gla.ac.uk CC: David Waters , Frank Wuerthwein , Jeff Tseng , Terry Watts , ALANSILL@tthep2.phys.ttu.edu, Flavia Donno Subject: Re: Database manpower References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I agree with you Rick. We have to tap into GRID resources. And as I said I will file just one big complain a year. So that is done and now I am cooperative. I cut the mailing list to begin with. And no hard feeling about Ombretta. Not your fault. As you said, she came at a bad time, with strings attached, and as I said, that's all the help I could find and clearly is not enough. I wander if it may be a good idea to try to investigate the EDG data replica and data management tools to describe a local data set, i.e. use the current EDG stuff to describe the local list of files. The idea is that somehow one should be able to send a job to the GRID broker that says "I want the j/psi version N", and some "place" says "I have that", by data set name, not by looking up file by file or object by object. The grid should provide the means for different sites to keep their own DB of which files are there with whatever DB mechanism they have/want, and exchange that kind of requests plus providing the full list of file names if you want to be sure, but only on request. Then locally I somehow extract the list of files from the DB and more or less like the FKW submitter does now on the CAF prototype pass the list to the AC++ exe. The nice fature is that I import the data whichever way it works, and use whatever file tracking software my sys.man. likes. Is this something reasonable ? If so, it might be small enough, and nuclear enough, that we can try to have a look at it with the limited italian resources. If someone thinks this is useless, or stupid, or a waste of time, I'll be thankfull for being educated before we start. This naturally then leads to the idea of "data set assurance", i.e. proving I really have the correct data set I claim I have, for this I had tossed that idea of data set checksum, and would not mind pursueing it. Stefano