How I Transferred a 2TB Backblaze B2 Snapshot to AWS Glacier for $2

If you’re storing large backups in Backblaze B2 but want to archive them long-term into AWS S3 Glacier Deep Archive, this guide will walk you through exactly how I did that — for just around $2.

---

🗺️ Why Move from Backblaze to Glacier?

• ✅ Glacier Deep Archive is the cheapest long-term cloud storage on AWS
• ✅ Redundant and geographically distinct storage option
• ✅ Archiving large B2 snapshots without downloading to your local machine

---

⚠️ Region Matters: Use us-west-1

Before spinning up any AWS resources, make sure your AWS resources are in the same region as your Backblaze B2 account.
In my case, both were located in us-west-001, so I used the us-west-1 AWS region to keep latency (and costs) down.

---

🧰 What You’ll Need

• An EC2 instance (no block storage needed)
• An S3 bucket with Glacier Deep Archive storage class
• Rclone installed
• Access credentials for both B2 and AWS

• You have already taken a snapshot of your B2 bucket via the Backblaze B2 web portal

🏗️ Setup Summary

1. Create a t3.medium EC2 instance in us-west-1 with Amazon Linux 2
2. SSH into it and install rclone
3. Configure rclone with both B2 and S3 remotes:

    rclone config
   

   in this example, I named B2 as b2remote and S3 as awsremote. Make sure you choose Glacier Deep Archine when configuring the S3 remote

4. Launch the transfer in a tmux session:

    tmux new -s b2transfer          
    rclone copy b2remote:your-bucket-name/your-snapshot.zip awsremote:your-glacier-bucket --progress --transfers=8  --checkers=16  --fast-list  --ignore-times  --size-only
   

5. This transfer might take a while, so to make sure the transfer is running in the background, do the following: Ctrl + B, then D.
6. You can now log out from the EC2 instance. You can revisit the progress later on via tmux attach -t b2transfer

7. To check whether the transfer is already done, just visit the S3 bucket and check if the zip file is already there

⏱️ Transfer Time & Performance

Transferring 1.9 TB of data took:

⏱️ 14h 46m 48.6s at an average of ~47 MB/s

Despite using the same region, there were notable network patterns:

• First hour: ~10 GB/hr transfer rate
• Remaining hours: dropped significantly to ~1 GB/hr

This may suggest throttling or region-crossing backend traffic between B2 and S3.

---

💸 Total Cost Breakdown

Resource	Cost	Notes
EC2 compute	~$2.00	t3.medium (14h runtime, no block storage)
S3 storage	Varies	Glacier Deep Archive is ~$1.00/TB/month

💡 No EBS volume was attached — I streamed directly with rclone

---

🧼 Cleanup Steps

Once done, don’t forget to:

• Stop or terminate your EC2 instance
• Delete any temporary files or logs
• Optionally, delete the original B2 snapshot if archived

---

🧪 Optional: Verify Integrity

rclone is reliable, but it’s good practice to verify your upload later:

```bash rclone md5sum b2remote:your-snapshot.zip > original.md5 rclone md5sum awsremote:your-snapshot.zip > glacier.md5 diff original.md5 glacier.md5

---

• ← Previous Post