Backups
Overview
Dragonfly Cloud supports manual on-demand backups and recurring scheduled backups. To ease the process of migrating data, importing Redis backups (RDB files) from a remote storage system is also supported.
Existing backups in your Dragonfly Cloud account can be restored to an active data store at any time.
For clustered (multi-shard) data stores, please review the cluster backup limitations before restoring.
Manual On-Demand Backups
To create a manual on-demand backup, follow the steps below:
- Click the three-dot menu () in the data store row.
- Expand the Backups menu and select Manual, the backup drawer will open on the right.
- Enter a meaningful Description that will help you identify the backup later for restoration.
- Set the Retention period to the number of days after which the backup will be deleted automatically.
- Click Create.
View created backups under the Account > Backups tab.
Scheduled Backups
To set a recurring scheduled backup, follow the steps below:
- Click the three-dot menu () in the data store row.
- Expand the Backups menu and select Schedule, the backup drawer will open on the right.
- Set the Backup Policy to Enabled, the Schedule options will appear.
- Select the day(s) of the week (or Every Day) and hours of the day (or Every Hour) to set when the scheduled backup should be created.
- Set the Retention period to the number of days after which the backup will be deleted automatically.
- Click Apply.
You can edit the backup schedule anytime.
Restoring from Backup
To restore a backup in an existing data store, follow the steps below, with caution:
- CAUTION: This cannot be undone. It will clear all data in the data store and replace it with the backup.
- Click the three-dot menu () in the row of the data store you would like to restore to.
- The Restore Backup drawer will open on the right.
- Make 100% sure you select the correct backup from the dropdown.
- Click Restore.
Cluster Datastore Backups
Dragonfly Cloud supports backups for clustered (multi-shard) data stores. A cluster backup captures the full dataset distributed across all shards at the point in time the backup was taken.
Limitations
- Shard count must match on restore: The target data store must have the same number of shards as the data store at the time the backup was created. Restoring a cluster backup to a data store with a different shard count is not supported.
- Shard distribution must be identical: Data is partitioned across shards by hash slot ranges. Because each shard's backup encodes data bound to a specific slot range, the slot layout of the target data store must exactly match the original. A mismatch will result in a restore failure.
- Cross-tier restore is not supported: A backup taken from a clustered data store cannot be restored to a single-shard (non-clustered) data store, and vice versa.
CAUTION: Before initiating a restore, verify that the target data store's shard count and configuration match the data store that the backup was originally taken from.
Viewing and Deleting Backups
- You can view all backups under the Account > Backups tab.
- To delete a backup, click the three-dot menu () in the backup row and click Delete.
Importing Redis Backups (RDB)
You can import Redis backups (RDB files) to Dragonfly Cloud from a remote storage system. Once imported, the backup will be visible under the Account > Backups tab and will be available for restoration.
To import a Redis backup, follow the steps below:
- Navigate to the Account > Backups tab.
- Click on the +Import RDB button.
- Enter a meaningful Description that will help you identify the backup later for restoration.
- Set the Retention period to the number of days after which the imported backup will be deleted automatically.
- Enter the HTTP Source URL of the RDB file to import.
- Chose the Target Cloud to import to, must match the cloud of the data store you want to restore to.
- Chose the Target Region to import to, preferably match the region of the data store you want to restore to.
- Click Import.
- Once the backup is visible in status Ready, you can restore it to a data store as described above.
RDB File Accessibility
- The imported RDB file URL must be publicly accessible via HTTP(S) over the internet.
- It is most secure to use a signed URL with a short expiration time to import the RDB file.
- See AWS S3 presigned URLs or Google Cloud Storage signed URLs for more information.
- Otherwise, you can generate a long random URL and delete the file after import to ensure security.