Bookkeeping

From ecology
Revision as of 12:12, 2 June 2008 by Bart (talk | contribs) (FILES)
Jump to: navigation, search

Description of tables in BOOKKEEPING schema in the Flysafe database

FILES

The FILES table describes the status of all data-files which have been or need to be processed. This information is used by the bookkeeping service on the flysafe website

ID unique file id
filename unique file name
file_type Type of file ie MPR, ECMWF,
file_size Size of file in bytes
status Status of file ie new, processed, archived
filename_time Time associated with filename as timestamp.
time_in Timestamp when file was first registered
time_changed Timestamp of last status change

Flysafe Grid-Store Catalog

The following three tables are used to implement the Flysafe Grid-Store catalog. If files are being archived or copied the Grid-Storage, this is administratred in the following tables.

For Grid-Storage it is preferable to have a limited number of big files instead of many small files. Therefore files are first collected into bigger archives, which are then stored as archive to the Grid-Storage. This is also maintained in the tables below.

FILE_CATALOG

Every file to be stored gets an entry in the FILE_CATALOG table.

id ID for file
filename Unique filename
filetype Type of file (MPR, tar, etc)
path Original path of file on Flysafe server


ARCHIVE_FILE

Files can be collected together in archive (usually tar files). These archive files are also registered as files in the FILE_CATALOG table. The ARCHIVE_FILE table contains the information about which files are in what archive file.

archive_id Id of archive file in FILE_CATALOG table
file_id Id of file in archive in FILE_CATALOG table
archive_path Path of file in archive

FILE_REPLICA

The FILE_REPLICA table contains the actual storage location of the stored files. It is possible to have more replicas of the same file.

file_id References in file_catalog
replica_id replicanumber, ie 1 for first replica, 2 for second etc
surl SURL is the storage location.
stored timestamp with time of storage
adler32 adler32 checksum of file
verified boolean indicator