Table of Contents:
Introduction
|
This tool is currently in development, and will provide a web-based UI for performing backups and restores to/from DuraCloud.
|
Scope:
- Any (user-selectable) set of objects or datastreams stored in Fedora.
- Any externally-managed ("E") datastreams available via http or file URLs.
Low-level storage options supported:
- All (interacts with Fedora via its APIs)
Status
In-development. The initial goal is to have a deliverable web application with basic selective backup and restore functionality in time for OR11.
Requirements
- Web-based utility that does not require a sysadmin to use
- Has a REST API for invoking from other services/UIs
- Backups to DuraCloud:
- Frequency: On-demand and schedulable
- Selection:
- User-supplied PID(s)
- User-supplied queries involving date ranges and other datastream/object attributes (e.g. custom "ready for sync" flags)
- Content: Fedora object XML, inline datastreams, managed datastreams, and external datastreams
- Granularity: Object-level, but provides a flexible way to filter out datastreams/datastream versions.
- Restores from DuraCloud:
- Frequency: On-demand
- Selection:
- User-supplied PID(s)
- By date range
- By DuraCloud item path
- Via RDF search of Fedora repository (when available)
- Provenance:
- Provides a standard way to record restore events, or an option to hook into a messaging queue so that they may be recorded as required by the institution (via a custom PREMIS datastream, etc.).
- Fixity
- Report on differences between what's stored in Fedora and in DuraCloud
- Automatic and manual repair of Fedora-DuraCloud inconsistencies
- Report on Fedora-Fedora inconsistencies (FOXML vs computed, as a built-in service?)
- Report on DuraCloud-DuraCloud inconsistencies (Stored vs computed, via fixity service)
More TBD - extrapolating from discussions and meeting notes thus far
UI Design
TBD - basic screen mock-ups
Related Meeting Notes