[ Home | Contents | Search | Post | Reply | Next | Previous | Up ]
From: Marco.Cattaneo@cern.ch
Date: 5/18/00
Time: 9:15:26 AM
Remote Name: 137.138.142.33
This is a multi-part message in MIME format. --------------274E3408F8DCE8C4E771C05B Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Dear colleagues, I have been asked to prepare the LHCb answers to a questionnaire concerning the proposed decommissioning of RISC Unix platforms at CERN. The answers from all the experiments will be collated and presented at a meeting of the FOCUS committee on 8th June, where it is expected to approve a decommissioning schedule. I attach my draft answers. Please let me have your comments and corrections by Friday 27th May at the latest. -- Marco Cattaneo | mailto:marco.cattaneo@cern.ch CERN, EP-ALC Group | http://cern.ch/~cattanem CH-1211 Geneve 23 | Tel: +41 22 767 4046 Fax: +41 22 767 9425 --------------274E3408F8DCE8C4E771C05B Content-Type: text/plain; charset=us-ascii; name="LHCb.txt" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="LHCb.txt" > > 1. How does your experiment define "frozen OS" support? (It is essential > that all LEP experiments answer this question. For other experuiments it > is optional) > Frozen support means the possibility of running existing experiment software without modification (e.g. no changes to compiler, libraries, file system, interface to the stager and to tapes etc.). Services used by the experiment, such as the stager, LSF etc. should be maintained in such a way as to allow the experiment software to continue working. Essential bug fixes should be applied to the CERN software libraries, but no new features are requested. > 2. Does your experiment agree with the RISC decommissioning timetable > presented by Manuel Delfino at FOCUS 17? (In summary, this was: RSPLUS/RSBATCH - No physicist access in 2001, removal end 2001 HPPLUS - Freeze OS end 2001, removal end 2003 CERNLIB - Last maintenance release early 2002 PAW replacement in production mid-2001) > LHCb sees no major problems with the decommissioning schedule of the RISC hardware: our software is currently supported on Linux, WNT and AIX. We have essentially stopped accessing RSPLUS/RSBATCH for production but still use it for debugging of FORTRAN code, due to the superior quality of the debugger compared to ddd on Linux. Concerning CERNLIB, we require no new functionality, but request that the current version be compiled for future versions of Linux and Windows OS/compilers installed on CERN public machines. While we are actively migrating our software to C++ and expect to have completed much of the migration by the end of 2001, we cannot say now when the last line of FORTRAN will have been removed from our production software. We request support for PAW until an OO alternative with similar (or enhanced) functionality has become widely accepted by the physics community and is officially supported by IT division.> 3. What are your experiment's requirements until 2005 concerning > import/export of data? > Currently we import MonteCarlo data on DLT 7000 from Rutherford and on Redwood from Lyon, as well as via the network. We expect that, in future, transfers via the network will increase in importance. Our current plan is to produce the following volumes of MonteCarlo data in the homelabs, which will be transferred to CERN preferably by network: 2000 0.5 TB 2001 0.6 TB 2002 0.9 TB 2003 1.7 TB 2004 3.5 TB Please note that the numbers are very provisional and may well be underestimates. In 2005 these numbers increase by two orders of magnitude in both directions (import of simulated data and export of real data) > 4. How critical is archiving for your experiment? > We make no use of pubarch. Some users archive their data on tapes, others in HSM. It is important that these data are safeguarded in any future migration. > 5. What is the 3-year requirement for remote backup? > We do not use remote backup for any of our systems. --------------274E3408F8DCE8C4E771C05B--