Deprecated

Backups are now supported natively by the D1 HTTP API. Refer to cloudflare/workers-sdk#5425 for an early wrangler version and information on the API endpoint. This package will no longer be maintained by me.

You can continue to use the package but since it has served its purpose, I will no longer continue to maintain it. The package is regardless in a stable state and there's no hurry in migrating to the Cloudflare API.

cloudflare-d1-backups

This is a package based heavily on sqlite-dump for Python by simonw but modified to be suitable for the current state of the open alpha, experimental backend (beta), databases.

See Cretezy's fork for a direct D1 HTTP API version.

This script has been tested on a 50mb+ database.

npm i @nora-soderlund/cloudflare-d1-backups

Create a backup

Creating a backup creates a SQL file in the destination R2 bucket.

• Install the package: npm install @nora-soderlund/cloudflare-d1-backups.
• Call createBackup(D1Database, R2Bucket, CreateBackupOptions | undefined) from an execution context.

CreateBackupOptions

• fileName: the file name, or a function that returns a string, for the SQL file in the R2 bucket, default is backups/${(new Date()).toUTCString()}.sql.
• maxBodySize: the maximum body size for R2 bucket uploads in MB, this depends on your plan, unless you're using Enterprise with a custom size, leave this unset.
• queryLimit: the row offset limit for each SELECT query, defaults to 1000. Alter depending on your row size.
• excludeTablesData: name of the tables from which you do not want to include data, ['example1'] (Array)
• tableNames: optional array of strings as table names, case sensitive

Restoring a backup

Assuming you've downloaded the backed up SQL file from your R2 bucket: npx wrangler d1 execute <database> --file=<backup.sql>

Example

See /example for an example worker.