I Tried Direct Access to Backup Data Using the New "Backup Access Points" Feature of AWS Backup for S3

I Tried Direct Access to Backup Data Using the New "Backup Access Points" Feature of AWS Backup for S3

AWS Backup for S3 now offers backup access points. Instead of running a restore job to retrieve a single file, you can fetch it directly through standard S3 APIs. I tested it with the AWS CLI.
2026.08.07

This page has been translated by machine translation. View original

Introduction

On August 6, 2026, an access point feature was added to AWS Backup for S3.

https://aws.amazon.com/blogs/storage/access-amazon-s3-backup-data-directly-using-s3-access-points-in-aws-backup/

Previously, even retrieving a single file from a backup required running a restore job. With the new feature, you can create an access point for a backup and access it directly.

Traditional restore Via access point
Run a restore job Create an access point
Need to wait for the job to complete Immediately accessible *1
Need to specify a destination bucket No destination bucket needed (direct download)
Retrieval per restore job Selectively retrieve any object

*1 Accessible as long as the backup is in the completed state (AVAILABLE)

In this article, we verify that you can create an access point with the AWS CLI and access backup data directly.

Verification

Environment

  • AWS CLI 2.36.18 or later (supports the create-backup-access-point command)
  • Run with administrator privileges

Preparing the S3 bucket

Create a test bucket and enable versioning (AWS Backup for S3 targets buckets with versioning enabled).

# Create bucket
aws s3api create-bucket \
  --bucket backup-access-point-test-YYYYMMDD \
  --create-bucket-configuration '{"LocationConstraint": "ap-northeast-1"}' \
  --region ap-northeast-1

# Enable versioning
aws s3api put-bucket-versioning \
  --bucket backup-access-point-test-YYYYMMDD \
  --versioning-configuration '{"Status": "Enabled"}' \
  --region ap-northeast-1

Place the test data.

for i in $(seq 1 10); do
  echo "Test content for file $i - created at $(date -Iseconds)" | \
    aws s3 cp - "s3://backup-access-point-test-YYYYMMDD/test-data/file-$(printf '%02d' $i).txt" \
    --region ap-northeast-1
done

AWS Backup configuration

Create a backup vault.

aws backup create-backup-vault \
  --backup-vault-name s3-access-point-test-vault \
  --region ap-northeast-1

Create an IAM role and attach the AWS managed policies.

# Create role
aws iam create-role \
  --role-name AWSBackupS3TestRole \
  --assume-role-policy-document '{
    "Version": "2012-10-17",
    "Statement": [{
      "Effect": "Allow",
      "Principal": {"Service": "backup.amazonaws.com"},
      "Action": "sts:AssumeRole"
    }]
  }'

# Attach policies
aws iam attach-role-policy \
  --role-name AWSBackupS3TestRole \
  --policy-arn arn:aws:iam::aws:policy/AWSBackupServiceRolePolicyForS3Backup

aws iam attach-role-policy \
  --role-name AWSBackupS3TestRole \
  --policy-arn arn:aws:iam::aws:policy/AWSBackupServiceRolePolicyForS3Restore

Enable the S3 resource type and start an on-demand backup.

# Enable S3 resource type
aws backup update-region-settings \
  --resource-type-opt-in-preference '{"S3": true}' \
  --region ap-northeast-1

# Start on-demand backup
aws backup start-backup-job \
  --backup-vault-name s3-access-point-test-vault \
  --iam-role-arn arn:aws:iam::xxxxxxxxxxxx:role/AWSBackupS3TestRole \
  --resource-arn arn:aws:s3:::backup-access-point-test-YYYYMMDD \
  --region ap-northeast-1
{
    "BackupJobId": "1dddfe69-5d32-4eee-a1bb-77bc75b8a7e0",
    "CreationDate": "2026-08-07T16:41:24.596000+09:00",
    "IsParent": false
}

After the backup job completes, obtain the RecoveryPointArn.

aws backup describe-backup-job \
  --backup-job-id 1dddfe69-5d32-4eee-a1bb-77bc75b8a7e0 \
  --region ap-northeast-1
{
    "State": "COMPLETED",
    "RecoveryPointArn": "arn:aws:backup:ap-northeast-1:xxxxxxxxxxxx:recovery-point:backup-access-point-test-YYYYMMDD-20260807074125-1de2a2cd"
}

Creating the access point

Create an access point for the recovery point (the actual backup entity). You can create up to 5 access points per recovery point. In addition, by setting an access point policy, you can restrict access by IAM principal, VPC, or object prefix.

aws backup create-backup-access-point \
  --name test-access-point \
  --recovery-point-arn "arn:aws:backup:ap-northeast-1:xxxxxxxxxxxx:recovery-point:backup-access-point-test-YYYYMMDD-20260807074125-1de2a2cd" \
  --region ap-northeast-1
{
    "AccessPointArn": "arn:aws:backup:ap-northeast-1:xxxxxxxxxxxx:accesspoint/test-access-point",
    "Status": "CREATING"
}

Confirm that the status has become AVAILABLE.

aws backup describe-backup-access-point \
  --access-point-arn "arn:aws:backup:ap-northeast-1:xxxxxxxxxxxx:accesspoint/test-access-point" \
  --region ap-northeast-1
{
    "AccessPointArn": "arn:aws:backup:ap-northeast-1:xxxxxxxxxxxx:accesspoint/test-access-point",
    "AccessPointMetadata": {
        "S3AccessPointAlias": "example-alias-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx-ext-s3alias"
    },
    "Status": "AVAILABLE",
    "BackupVaultName": "s3-access-point-test-vault",
    "ResourceType": "S3"
}

Direct access to backup data

Specify the value of S3AccessPointAlias as the bucket name and access it with the standard S3 API. Access via the access point is read-only (this article verifies only read APIs).

ListObjectsV2 (listing objects)

aws s3api list-objects-v2 \
  --bucket "example-alias-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx-ext-s3alias" \
  --region ap-northeast-1
{
    "Contents": [
        {
            "Key": "test-data/file-01.txt",
            "LastModified": "2026-08-07T07:40:15+00:00",
            "Size": 63,
            "StorageClass": "AWS_BACKUP_WARM"
        }
    ]
}

file-02.txt through file-10.txt could be retrieved in the same way, and all had Size 63 and StorageClass AWS_BACKUP_WARM. This is the value returned for AWS Backup backup data.

HeadObject (retrieving metadata)

aws s3api head-object \
  --bucket "example-alias-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx-ext-s3alias" \
  --key "test-data/file-01.txt" \
  --region ap-northeast-1
{
    "ContentLength": 63,
    "ETag": "\"fdef648f5a8cf52454ed8db9c21559b6\"",
    "VersionId": "Kjfj_zH.wbVvhruWZnprEe8RBRYd10Wg",
    "ServerSideEncryption": "aws:backup",
    "StorageClass": "AWS_BACKUP_WARM"
}

ServerSideEncryption was set to aws:backup (encryption managed by AWS Backup).

GetObject (retrieving file content)

aws s3api get-object \
  --bucket "example-alias-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx-ext-s3alias" \
  --key "test-data/file-01.txt" \
  /tmp/backup-file-01.txt \
  --region ap-northeast-1
$ cat /tmp/backup-file-01.txt
Test content for file 1 - created at 2026-08-07T16:40:14+09:00

We were able to retrieve the backup data directly without running a restore job.

Summary

The access point feature of AWS Backup for S3 makes file-level access to backup data easy. Once you create an access point, you can access the data directly with the standard S3 API (ListObjectsV2, HeadObject, GetObject) without running a restore job or preparing a destination bucket.

It looks useful for use cases such as "I want to recover just one file I accidentally deleted" or "I want to check the contents of a backup." The official blog also introduces compliance audits and forensic investigations as effective use cases.

Share this article

AWSのお困り事はクラスメソッドへ