Disk Storage and Datasets at SLAC

The general dataset access tool of dq2 is described in the usage of dq2. A very useful tool is the dq2 dataset browser allowing examine available data from different generators and releases. For problems in dq2 usage, the dq2-DDM operations savannah page has links for various related info, in particular the bug tracker in development tools section seemed to be the place user problems are reported and addressed.  Some specific instructions for pulling datasets to SLAC is also available.

Local ATLAS Disks at SLAC:

The larger new Tier-2 user disks are all managed through XROOTD which is the preferred location for all user datasets.  Please take a look at the instructions below for the usage of the XROOTD space.

• Using XROOTD

There are also some smaller old disk spaces (close to be full) which are intended as scratch area for releases, log files and smaller nutpules etc. together with some historical old datasets:

Note: Again, for normal datasets you should think about the XROOTD space first.  Any dataset you wish to move into the above disks will be sharing with user working space which are spread out on the last 4 disks, and the dq2 operations space on the two disks indicated by the last column of the table. So it is important to take a look at the space left on the disk you are targeting for storing samples (with command df -k .) to make sure it won't be close to full, before shipping a large sample.  ATLROOT is /afs/slac.stanford.edu/g/atlas/ at SLAC.

New datasets at SLAC, including pileup datasets

These datasets reside in the XROOTD storage element at SLAC. In order to use XROOTD, please take a look at the instructions below. Other useful links are provided at the top of this page.

• Using XROOTD

I have also slightly separated the pileup samples reconstructed with Release 12 from those done with Release 13. The differences are non-trivial, both in the EDM itself as well as in the pileup digitization and reconstruction. A discussion is ongoing as to some subtle yet perhaps very important differences in the pileup digitization, as can be read about in this meeting:

• Pile-up Meeting on 7 December 2007

Old CSC11 datasets at SLAC
 

Last Update:  Jan/10/2008
David Miller