Scholars' Mine

Backup Strategy

1. Repository databases have real-time redundancy that runs continuously, and full nightly backups are kept of the entire database.
2. Storage servers for production files have full hardware redundancy. It is protocol to create hourly and daily snapshots and offsite backups of production file systems. In this case, ‘offsite’ means offsite from the service provider’s co-location facility. Production files are mirrored every half hour, databases every 2 hours. A second copy of offsite backups is sent to Amazon S3. (Amazon S3 data is redundantly stored across multiple Amazon facilities and multiple devices in each Amazon facility.)
3. All uploaded files are stored in triplicate in a redundant storage cluster, as well as backed up offsite to a third-party cloud service, Amazon Glacier, that specializes in data archiving and backup. Glacier performs regular, systematic data integrity checks and is built to be automatically self-healing. 
4. Copies of the repository are also pushed to the Missouri S&T’s Amazon S3 account. New content and revisions to existing content are pushed automatically to S3.
5. Pre-production files and other source materials are maintained on storage servers provided by campus IT. These servers have real-time redundancy that runs continuously, and full nightly rotating backups are kept of the entire database. A second copy of these materials is maintained in Google Drive.

Data Recovery Provisions

All production servers are maintained at a high availability co-location facility with multiple backbone connections and backup generators or in redundant Amazon Web Services availability zones. We maintain failover web, database, and storage servers to continue to serve content in case of failures. In the event of a natural disaster or other catastrophe, the repository and its data can be restored from the backups.