Types of data in LHCb

Policy to store the data

• All express stream data stays on disk at CERN for the entire year
• The fraction of RAW and SDST processed at CERN (~20%) stays on disk at CERN throughout the entire year. The 80% processed at T1 sites does not, at least not at CERN (I don't know what the policies are at T1). Note that this also means 20% of runs at CERN, 80% at T1, since complete runs are processed at a give site.
• The 80% of RAW and SDST not processed at CERN cannot be staged at CERN by "normal" users.
• Tools to allow users to identify the runs on disk at a given site are being developed

All of the above is done to optimise the usage of disk space and to avoid breaking the staging system. It is nothing new, it is all documented in the computing model.

Comment: this is very important for the calibrations, as they are done on the express streams and they need the raw data! -- ElisaLanciotti - 28-Oct-2010