Administration

This page describes the organization of data in the Archival Storage so that the administrator can explore and debug the data.

Workspace

• Objects are stored as temporary files to the workspace of the Archival Storage during both upload and download requests. The temporary file is named with UUID which does not correlate with any ID of any object.
• Path to the workspace folder may be configured in arcstorage.tmp-folder entry of application.yml file.

Logical Storages

• The location of the object at logical storage depends on the dataSpace attribute of user which has uploaded the object.
• ID of the SIP (=ID of the AIP) and ID of general object at the storage is the same as the id in the database. ID of the ARCLib XML object at the storage is different from its ID in the database and is in following form: idOfTheSIP**_xml_**version.
• Logical storages also stores some metadata which might be used to recover database if its lost.

FS/ZFS

In FS/ZFS the data are stored in the folder structure depending on

1. rootDirPath entry of JSON string of config attribute of the logical storage
2. dataSpace of the uploading user
3. ID of the object

For example, object with ID cc5b7281-d8de-4f0a-8c06-6215c5359d72 of the user with dataSpace arclib at storage which has its rootDirPath set to /opt/archival-storage/data will be located in folder: /opt/archival-storage/data/arclib/cc/5b/72

Storing metadata:

• initial checksum of the object: stored during creation to the same directory as file into text file with file name and .{@link ChecksumType} suffix
• creation time of the object: provided by filesystem
• state of object: handled by empty file with original file name and .STATE suffix (e.g. fileId.PROCESSING)

Ceph

In Ceph the data are stored in the bucket of the same name as the dataSpace attribute of the uploading user.

Storing metadata: Metadata are stored in a separate metadata object, which id is idOfTheObject**.meta**