Snapshot vs Backup
Snapshots and Backups provide data protection for Virtual Machines (VMs). Both methods can be used to restore a full VM to a known state. So how do you know which is the best choice when requesting data protection for your application?
Snapshots
- Snapshots are recommended for situations where you need to quickly fall back to a known state such as before a new application install, upgrade, or patching.
- Snapshots are usually taken immediately before a change/update takes place.
- Snapshots can only be used for full VM restores or individual hard drives. Snapshots cannot be used for file level restores.
- Snapshots “freeze” the disk image for the VM and then track all changes to a snapshot chain.
- This takes up disk space on the hypervisor and over time can lead to performance degradation due to snapshot chain growth.
- Snapshots are not recommended for long term retention. Users will be given the option to retain snapshots for 7 or 14 days.
Backups
- Backups are meant for long term retention and run on a daily schedule. Most VMs are backed up every 12 hours, or 2 times per day.
- Like Snapshots, Backups can be used to restore full VM images and VM hard drives.
- Backups can also be used to search for and restore individual files/folders.
- Backups are saved to local storage on the Rubrik appliance in ARDC. When the backup is complete, it is immediately archived to Glacier Instant Access in AWS.
- The local copy is kept onsite for 14 days, which allows for rapid restores. A typical restore will take 15-30 minutes, but in an application down situation we can restore a copy directly on the Rubrik appliance in less than 5 minutes.
- OIT will have to move the VM off the Rubrik appliance at a later date.
- After 14 days, we can restore from the archive. The archived copy is kept for 95 days.
- Restores from archive take longer because Rubrik has to download the information from AWS first. Restores from archive typically take 45 minutes to 2 hours to complete.