Kanga Data Distribution - Import, Backup and Deletion
Tools
This page introduce to BBRORA installation
on remote sites and will explain the advantages to use it for import and
backup jobs of kanga files.
What is BBRORA ?
BBRORA (BaBaR ORAcle) is one of the
Oracle database at SLAC. Data files production information (sizes, analysis
group, processing releases, files path and name and more) is stored in
it. skimData is the user tool to access this information at SLAC.
The aim is to mirror this database on a local site that hasn't Oracle
engine, but a simpler one, like MySQL (above all, FREE!). Why do that ?
It should be desirable to have the possibility to use skimData locally
rather at SLAC, if locally there also are kanga data (or part or them).
The second advantage is the simplification of import and backup jobs.
At the moment some sites syncronize their data with SLAC using several
rsync-based tool. Often rsync shows too long times during the preliminary
phase (retrieving the list of files to import). Instead updating the local
BBRORA database is much faster. Once the BBRORA database is updated locally,
it can be queried (very quickly) to get a list of files that aren't present
yet, and import them. Another interesting task is to manage a backup system
synchronozed with the same database.
Managing data retrieving with local skimData and BBRORA
Once BBRORA is correctly installed at remote site skimData can be used
exactly in the same way it was did at SLAC. There's an additional information
for each entry (kanga file) in the mirrored BBRORA database. This information
is contained in the `import_status' field.
In
this
document it's explained much more about this field, but for now take
into account that it can assume the values: 0, 1, -1, 2, 0B, B, E; each
value corresponds to a physical position of the relative file (1="not yet
on disk", 0="on disk", 0B="on disk and on tape", etc. etc.). So if you
want to gather informations on entries (kanga file) with
import_status=X
do:
skimData --stream <DESIRED_SKIM/STREAM>
-g <DESIRED_RUNS> <OTHER_OPTIONS> --remote X
Generally it's not needed --remote option,
and skimData will work fine anyway (because it will return all
entries tagged as `0' or `0B', that mean "on disk" or "on disk and on tape".
This means that the users, that do physical analysis, have to use skimData
without --remote option, because they want
fetch the already imported files.
A set of tools to create, update and mantain BBRORA, and to import,
backup and delete kanga files, using the information in it, has been added
to SkimTools package. This package contains also a new version of skimData
that can run on remote sites and that recognize the --remote option.
Page written by Alvise Dorigo
Alvise.Dorigo@pd.infn.it
Web Page
Last update: 6 June 2001.
|
![[new]](new.gif){kind=small}