How Can We Help?
Disconnecting an external repository (e.g. DSpace)Disconnecting an external repository (e.g. DSpace)
Inform Elsevier before uploading more than 100 GB of data to your Pure in any one week. If you need support from an Implementation Manager to disconnect an external repository from your Pure, please contact your Customer Success Manager.
The process of disconnecting an external repository and transferring all files from the repository to Pure's own storage involves several steps. It is recommended that you test the whole process with a staging Pure and a staging version of your external repository, as local configurations of both systems can influence the results.
Process
1. Back up Pure and your external repository
Create backups of both systems in case of unexpected results. If your Pure is hosted by Elsevier, backups of it are automated. See Backup.
2. Update Handle.Net (if in use)
If links to files in your external repository are managed with Handle.Net, the handles must be updated with the links to these files in the Pure Portal before you disconnect the repository. The Pure Handle.Net plugin can be used for this, after some preparation work.
- Add a new secondary source ID to each record in Pure that has an external repository ID, using the External Publication ID Import job. The source will be ‘handle.net’ and the ID itself will be the handle identifier (this might be the same as the external repository ID, but it does not have to be).
- Install and configure the Pure handle.net plugin, which will add the Pure Portal links to the existing handles in Handle.Net. See Handle.Net.
3. Ensure you have enough storage in your Pure
Check the total size of the files in the external repository that you will be transferring to Pure. If your Pure is hosted by Elsevier, 1TB is included; if you need to purchase more, contact your Customer Success Manager. See Storage.
4. Remove the conditions for storage
Go to Administrator > Storage > File storage and click Edit next to the external repository store. Uncheck all of the conditions for storage.
5. Run the Preserved Content Update job
The next time the scheduled Preserved Content Update job runs, all the files will be downloaded from the repository and saved in Pure’s storage. The job can also be triggered manually in the job configuration by toggling on fileSynchronizationJob.performFullScan.
6. Check the storage is empty and delete
Check the Preserved Content Update job log, as well as Administrator > Storage > Content status, for warnings or errors and resolve them. See the FAQ below for common errors and how to resolve them. Then go to Administrator > Storage > File storage and check the external repository store is empty. Click Delete to remove it.
FAQ
How long will the Preserved Content Update job take?
This is difficult to estimate. It is recommended that you test it on a staging system.
Is any downtime required?
No. The Preserved Content Update job is designed to be run on a regular basis without impacting the performance of your Pure.
Can the connection to the repository be temporarily set to read-only?
Yes. The 'read-only' configuration will temporarily prevent Pure from adding files to a repository if, for instance, the repository is down for maintenance. It is not intended to be left on for extended periods of time, and should not be used as an alternative to disconnecting a repository. The feature works by disabling the connection between Pure and the repository, and not the job that moves files to the repository; Pure will continue to try to write to the repository and will log errors when it cannot connect. When a file is changed in Pure, it will be put in temporary storage and Pure will start trying to move it to the repository. Since it cannot connect, the file will not be moved to Pure’s local storage until read-only mode is disabled again.
Can separate stores be created within Pure's S3 storage?
No.
Will files moved from the external repository to Pure get a second handle in Handle.Net?
No. The Pure Portal links will be added to the existing handles in Handle.Net if you use the method described above.
Is it necessary to register a new handle prefix for Pure in Handle.Net?
You decide on the prefix to be used by Pure and configure it in Handle.Net; it could be the same one you used for the external repository.
How can files/metadata be removed from a repository store in Pure when the Preserved Content Update job cannot move them?
In some cases Pure is unable to move all repository files into Pure's default store. If, for example, a file, or even an entire record, has been previously deleted inside the repository, or other changes have been made that led to a record getting a new ID that deviates from the one stored inside Pure, Pure will lose the connection to that record/file (no changes should be made inside the repository for this reason). When the repository connector is disconnected, the Preserved Content Update job will be unable to fetch the affected records/files, and the entries for them will remain inside the repository store at Administrator > Storage > File storage. An overview of all failed processes can be seen at Administrator > Storage > Content status by activating the correct filters. An Implementation Manager can assist you with updating or removing the references to the repository connector on the affected Pure records, which will remove the entries from the Pure repository store. After it is completely empty, you can delete the Pure repository store. If you still have access to any missing file(s), they can be reattached to the relevant Pure records.
Why does a seemingly empty repository store in Pure not have a Delete button?
If the Pure repository store at Administrator > Storage > File storage shows ‘0 files stored. 0 metadata stored’ but does not have a Delete button, a script needs to be applied to remove the repository connector before the store can be deleted. An Implementation Manager can assist you with this.
Published at May 05, 2025