Workflows: Difference between revisions

Revision as of 17:22, 21 July 2021

Grids - overview of the computing farms and job submission
HPC - high performance computing resources
Docker software containers for submitting to special farms
analysis workflow - how to read existing datasets
production simulation workflow - how to run simulation, concatenate and upload the files
MARS and G4Beamline workflow - these are not in the art framework
Job planning - considerations of project CPU, disk, memory and file size
Prestage - make sure files have been retrieved off tape and are ready to use
GenerateFcl - generate fcl for simulation jobs
SubmitJobs - submit, monitor and recover jobs in the workflow
Concatenate - merging root and art files
Upload - upload files to tape
ErrorRecovery - how to recover from common errors
Validation - check how code is performing
Datasets - production datasets
RunNumbers - run number ranges and uses

dCache - the large aggregated data disk system
enstore - the mass storage tape system
Data transfer - how to move larger datasets, necessary for grid jobs
File names - how to name files for upload or for production
File families - the logical grouping of data on tapes
SAM - file metadata and data handling management system
- SAM metadata - the metadata stored for each file
- SAM expert - some notes for the future, not of general interest
File tools - tools for manipulating large file datasets and metadata
FTS Upload - deprecated method to upload by the FTS area
Rucio - next-gen data handling

@@ Line 11: / Line 11: @@
 * [[Prestage]] - make sure files have been retrieved off tape and are ready to use
 * [[GenerateFcl]] - generate fcl for simulation jobs
-* [[gridexport]] - export a build of Offline or a satellite release for use on the grid
 * [[SubmitJobs]] - submit, monitor and recover jobs in the workflow
 * [[Concatenate]] - merging root and art files