General scheme for using the Globus infrastructure at SUNY
The version of Sat Mar 8 19:49:00 EST 2003.
| Short explanation of general terms and scenario
(this explanation is true for concrete realization at SUNY) | |
|---|---|
| Actions | Short Description |
| The data files | If you need to know where the files with physics data are located please use File/Replica Catalog |
| Job submission | The submission of the script for execution. Job submission for distributed infrastructure could be done with commands in the table . |
| Site | the place where physics data are located. The site has name, for instance SUNY or PHENIX. |
| Location | the exact place at site where the physics data are located. For example they might be on local file system, NFS mounted file system, on HPSS, etc. Usually exact place means host and directory name where files with physics data are located. For example: rserevr1.i2net.sunysb.edu:/Data/data02/nnajit. |
| Collection | collection of files at a site. Collection might include all files in a location. |
| Replication | Exact copy of file or group of file in different location |
| Moving the data | If you need to move the data (currently from PHENIX to SUNY and back) you need to check the existing locations
(site, host and directory tree) in the Catalog. If it does exist already you might need
to create collection and replication task (please see Catalog). The replication task
is performed usually automatically during the day by special procedure started by cron daemon.
In general the scheme could copy only collection of files from source location to destination location. That means you have to describe both locations and the collection before creating the replication task. |
Updated: March 8, 2003
WebWork:
Nuclear Chemistry Group: Andrey Shevel