R2 Super Slurper reaches GA for one-time migrations
R2, Cloudflare's object storage platform, eliminates egress fees, and its S3-compatible API makes it straightforward to point existing applications at R2 buckets. The harder part has historically been moving existing object data into R2 from other providers. Cloudflare introduced R2 Super Slurper in beta to handle that job, and the migration tool is now generally available for one-time data transfers from S3.
During the beta period, early adopters used Super Slurper for hundreds of S3-to-R2 migrations. Cloudflare says that feedback drove several improvements to the tool: copy throughput is up to five times higher than at launch, reliability has been strengthened, and support was added for copying data between two R2 buckets. The GA release focuses on one-time migration, which covers most move-to-R2 scenarios.
Parallel, resumable transfers
Super Slurper copies objects from S3 in parallel and runs those transfers over Cloudflare's network to maximize available bandwidth. Migrations that would take a full day with other approaches have completed in roughly half an hour, according to a quote from PDQ Engineering Manager Nick Inhofe that Cloudflare included in its announcement.
Network transfers can fail, so Super Slurper includes retry logic to push multi-terabyte migrations to completion. Individual objects larger than a certain threshold are transferred in chunks, so a failure only requires retrying a portion of the transfer rather than the whole upload. That approach speeds up migrations and reduces cost. Objects that still fail to transfer are logged, allowing you to handle them after the migration finishes.
Kick off a migration from the dashboard
The migration workflow is entirely dashboard-driven:
- Under R2 in the Cloudflare dashboard, select Data Migration.
- Select Migrate files.
- Enter your Amazon S3 bucket name, an optional bucket prefix, and credentials for that bucket, then select Next.
- Enter your R2 bucket name and credentials, then select Next.
- Review the migration details and select Migrate files.
An asynchronous job then performs the copy, and you can follow its status from the dashboard. The same process works for moving data between R2 buckets; just select Cloudflare R2 as the source provider. Documentation for Super Slurper is available on the Cloudflare developers site.
Cloudflare singled out a testimonial from Reelgood Backend Engineering Manager Damien Capocchi, who said the tool allowed the company to migrate millions of images quickly — something that would have been impractical to build in-house at a reasonable cost.
What comes next
One-time migration is sufficient for most transitions from S3 to R2. But Cloudflare is looking to support a different pattern with an incremental migration feature that is still in the works: copying data to R2 "sip by sip" as it is requested.
The design goal for incremental migration is straightforward. When an application makes a request that the R2 bucket does not yet handle, the object is downloaded from the origin bucket and copied to R2 — likely for the last time, because that same request is then served by R2. Data migration then rides along normal application traffic, which means no extra egress fees beyond the final pulls. Switching an application's S3 endpoint to R2 would be all that's needed to begin the process. Cloudflare is running this as a private beta, and interested users can join the associated waitlist. The company also pointed to its Discord community for questions about R2 and related tools.
