Using Kanga Files at Remote Sites

Kanga and ROOT analysis environment

When this is installed at your site, you will be able to read Kanga (ROOT) files from Beta jobs running on your machine.

Local skimData database ("BBRORA")

This allows skimData to be run locally (otherwise you will have to use skimData to produce TCL files at SLAC, transfer them to your machine, perhaps adjusting for any differences in Kanga file availability - the skimDataRemote script may help with this, but is still rather cumbersome).

Local skimData uses an SQL database, which can be stored in MySQL (free software) or Oracle (so far, the Oracle version has only been used at SLAC). It is updated from SLAC using skimSqlMirror.

Automatic import of Kanga files

Files are selected for import with skimSqlSelect and copied using skimImport. These tools require a local skimData database.

Kanga files can also be transferred using syncslac and rsync, which doesn't require a local skimData database. It can make a copy of all or part of the Kanga directory tree. syncslac/rsync is the old method, but does still work (after a fashion).

Management of Kanga files

Files can be backed up or archived with skimBackup. Files can be removed with skimDelete. These tools require a local skimData database.

Kanga conditions data

Kanga conditions data is stored in $BFROOT/kanga/CondDB/ at SLAC. This should be copied separately (eg. using syncslac).