Improving DSpace Backups, Restores, Migrations

摘要

In the past, backing up your DSpace contents has involved semi-synchronized backups of both your database and your files. Although this procedure generally works fine, it can prove problematic when you suddenly need to restore the contents of a single Community, Collection or Item (both metadata and files). There is also the problem of metadata and content files residing in separate backups - if either one of these backups becomes corrupted, it is nearly impossible to completely restore your DSpace contents. This talk will introduce a new DSpace feature being developed as part of the DuraCloud integration project. This new feature will allow you to export your entire DSpace Community / Collection hierarchy (including all Items, and their metadata and files) into a series of METS-based packages. These METS-based packages may be used to restore all of your DSpace contents (into another DSpace), or just the contents of a single Community, Collection or Item. These packages can also provide a more stable way to backup your DSpace contents, or an additional means of getting content in or out of DSpace. This work is based on a DSpace plugin built as part of the MIT PLEDGE project; however it has been updated to allow for a complete restore of your DSpace hierarchy.