Versioning of files

With the start of provisioning for DataStorage.nrw resources, a new feature is also activated: File Versioning.

This works in a similar way to the Git commit history by retaining all file versions. If a file with the same name as an existing file is uploaded, the new file will be saved, but the old file will be kept as the previous version instead of being overwritten. There is no maximum number of file versions that are retained. File versioning in Coscine is designed to allow the recovery of deleted files after a short period of time and not to keep versioned files forever.

Note

Please note that file versions block storage space within your resource, even if the respective file versions are not displayed in the UI or API. Storage space blocked by deleted versions is freed up 7 days after deletion.

When using S3 clients, versions might be displayed in the interface, depending on the client. However a deletion is not possible and the versions will block the storage space until they are finally deleted.

Integration in the UI is planned for the future so that more individual settings for deleting versions will be possible.

Quota management - best practices to save space

Versions can become a huge factor when planning how much quota you need. Try to plan ahead how often you will update files.

Update frequency

If you update your whole data once per week, you constantly have two versions. If you update daily, you have seven versions up to eight, depending on the difference between reupload time versus the time of the scheduled deletion. Keep your update frequency reasonable

Folder reuploads

If you are regularly reuploading newer versions of files, do that individually and upload them into the folder to save your quota. A typical case how unnecessary versions are created is by reuploading folders, where only few files have been changed. This typical behavior is learnt cause it is convenient and works well with modern operating systems like Windows/Linux/Mac OS. Unless those systems, coscine does not yet offer functions for skipping overwriting operations for files with the same filename and size. In case whole folders are reuploaded, a new version is created for all files, which in the end occupies quota. Upload files individually, not whole folders