By default, Atlas Cloud Backup stores snapshots in the same cloud provider region as the cluster's current primary node. You can configure Atlas to store copies of your snapshots in additional regions by defining a snapshot copy policy for your cluster.
A snapshot copy policy includes one or more policy items. Each item defines the following for a specified cloud provider region:
Frequency for copying snapshots to that region
Retention time for snapshots copied at that frequency
Whether to copy oplogs to support point-in-time restore in that region
Snapshot copy policies can improve resiliency and restore speeds for your cluster. If your primary region goes down and your original snapshot becomes unavailable, Atlas can use copied snapshots in other regions to restore your cluster. Additionally, if you copy snapshots to every region in a multi-region cluster, Atlas can perform faster direct attach restores instead of slower streaming restores to restore your cluster after a regional outage.
If you enabled Continuous Cloud Backup for your cluster, you can also configure your snapshot copy policy to copy oplogs to your snapshot copy regions. Atlas can perform a PIT restore using a copied oplog in the event of a regional outage in your cloud provider.
Required Access
To modify or delete backup policies, you must have
Project Backup Manager or Project Owner
access to the project. Users with Organization Owner
access must add themselves as a Project Backup Manager or
Project Owner to the project before they can modify or delete
backup policies.
Considerations
If you have a Backup Compliance Policy enabled with the Keep all snapshots when removing additional snapshot regions option set to On and you modify or delete a snapshot policy item, any existing snapshots are retained until their scheduled expiration date.
If you change your cluster's primary region to a secondary region that you currently have configured for snapshot copy distribution, Atlas automatically updates your backup policy to distribute snapshots copies from your new primary region to the original primary region (now a secondary region) and other secondary regions configured for snapshot distribution.
If you enable a peering connection and restrict certain regions for Google Cloud, you can copy snapshots to only those regions.
Oplog copy distribution is enabled at the copy region level. If you set Point-in-Time Restore to On for a snapshot copy policy item, Atlas copies oplogs to that copy region and enables point-in-time restores for all snapshot copy policies in that copy region.
Enable Snapshot Copy Distribution
Update Snapshot Copy Policy
Disable Snapshot Copy Distribution
Example Snapshot Copy Policy
This example snapshot copy policy distributes copies of snapshots taken for the following backup policy:
Backup Policy:
This example backup policy takes hourly, daily, and weekly snapshots and stores them in the same region as the cluster's primary node.
Frequency Unit | Every | Retention Time | Snapshot Time |
|---|---|---|---|
Hourly Snapshot | 6 hours | 2 days | 18:30 UTC |
Daily Snapshot | N/A | 7 days | 18:30 UTC |
Weekly Snapshot | Monday | 4 weeks | 18:30 UTC |
Weekly Snapshot | Friday | 4 weeks | 18:30 UTC |
Snapshot Copy Policy:
This snapshot copy policy copies daily and hourly snapshots to the AWS
region us-east-2 (Ohio), and copies all weekly snapshots to
us-west-1 (N. California). It retains each snapshot copy in the copy
regions for the specified retention time, and copies oplogs to the
us-east-2 region to enable PIT restores.
Region | Snapshots | Retention Time | Point in Time Restore |
|---|---|---|---|
| Hourly | 12 hours | Enabled |
| Daily | 7 days | Enabled |
| Weekly | 7 days | Disabled |