AWS Snow Family - Runbook & Reference¶
中文 · GitHub source Facts verified against official AWS documentation: 2026-08-19
Overview¶
The AWS Snow Family provides physical devices for offline data transfer and edge computing in environments without reliable connectivity. Note the current lifecycle: Snowcone (HDD/SSD) was discontinued on November 12, 2024, and AWS Snowball Edge is no longer available to new customers. For new online data transfers, use AWS DataSync; for offline transfer options, review the current Snowball Edge documentation.
Devices and current status¶
| Device | Purpose | Status |
|---|---|---|
| Snowball Edge Storage Optimized (210 TB) | Large-scale offline data migration, S3/EC2-compatible endpoints | No longer available to new customers |
| Snowball Edge Compute Optimized | Edge compute with local processing | No longer available to new customers |
| Snowcone (HDD/SSD) | Small edge compute/data transfer | Discontinued November 12, 2024 |
| Snowmobile | Exabyte-scale data center transfer | Retired March 2024 |
Key concepts¶
- Job: an import/export job for moving data between your site and Amazon S3; create and manage jobs in the console or with the Snowball API.
- Device configuration: Storage Optimized (210 TB, up to 40 vCPUs) vs. Compute Optimized (up to 104 vCPUs with GPU options).
- Endpoints: Snowball Edge exposes S3- and EC2-compatible endpoints plus NFS, for local workloads.
- Clusters: group 3-16 devices for local storage and compute with higher durability.
- OpsHub / Snowball Edge client: tools for unlocking the device, configuring the network, and transferring data.
Common operations (AWS CLI)¶
# Create an import job (S3 destination)
aws snowball create-job --job-type IMPORT --resources '{"S3":{"S3ResourceArns":["arn:aws:s3:::bucket-name"]}}' \
--address-id <address-id> --role-arn arn:aws:iam::123456789012:role/snowball-role \
--snowball-capacity-preference T210 --shipping-option SECOND_DAY
# List and inspect jobs
aws snowball list-jobs
aws snowball describe-job --job-id <job-id>
# Update the shipping address or state
aws snowball update-job --job-id <job-id> --address-id <new-address-id>
# Cancel a job before the device ships
aws snowball cancel-job --job-id <job-id>
Best practices¶
- For ongoing or online migrations, use AWS DataSync instead of physical devices where possible.
- Estimate data volume and transfer time before ordering; choose the right device size to minimize shipping legs.
- Set up the S3 bucket, IAM role, and shipping address before creating the job.
- Use AWS OpsHub to unlock and monitor the device; keep the unlock code and manifest secure.
- Enable encryption: data on Snow devices is encrypted at rest and in transit by default.
- Track the device with the E Ink shipping label and follow import/export return procedures.
- For edge compute, snapshot AMIs and plan for device replacement; AWS monitors connected devices and can ship replacements.
Troubleshooting¶
| Symptom | Checks and fixes |
|---|---|
| Device won't unlock | Verify the job state and use the correct manifest/unlock code from the console. |
| Slow local transfer | Check the local network between clients and device (10/25/40/100 GbE), and use the S3 adapter or NFS as appropriate. |
| Data not appearing in S3 | Confirm the job completed and was processed by AWS after the device was returned. |
| Job cancelled after shipping | Contact AWS Support; shipped jobs generally cannot be cancelled. |
| Edge compute instances fail | Verify AMI compatibility with Snowball Edge (sbe1/sbe-c/sbe-g instance types). |
Limits¶
Job counts per account, device counts in flight, and cluster sizes (3-16 devices) are constrained by AWS quotas and regional availability. Check the Service Quotas console and Snowball Edge documentation for current values.