[[asc:workbook| <Please do not keep data on NFS where your home directory is (/users/). There is a significant performance penalty when running your jobs, plus it is impossible to backup your data To run over a large data sample, we need to copy data on the computer farm and then you can process the data using Arcond/Condor. Read [[asc:workbook_cpf|here]] more. Please contact S.Chekanov if you need to copy data on servers of the computer farm. Usually (but not always), we keep two copies of data: one on the central storage and second version distributed (staged) on the farm. Typically, you need about 1 day to process 1 TB of data. On the computer farm, it should take less than 30 min. How to get the data from the grid is described [[asc:workbook_grid|here]]. You can use use either dq2-get or subscription [[asc:workbook_grid_subscribe|Using data subscription]].