AWS S3 or S3-compatible

Overview

Storware Backup & Recovery can store backups in AWS S3 or S3-compatible backup providers. In most cases, you just need to prepare a bucket (with versioning enabled if possible) and generate an access/secret key for Storware Backup & Recovery. Storware Backup & Recovery can be installed in AWS (if EC2 backup is used), but in most cases, S3 is used just as a cloud backup provider for on-prem environments.

Typical use cases are:

• When AWS is used - choose a single bucket with versioning enabled - all backup objects will have names in /container_name/path/to/backup format, where container_name typically is the VM name with an identifier.

• When a 3rd party is used - you need to verify:

  • Which strategy is supported by the vendor - i.e. Scality requires a single bucket without versioning.

  • When timestamp recording of the object should occur - i.e. Scality does it after data is stored (unlike AWS).

Storware Backup & Recovery also support encryption. It uses server-side encryption with customer-provided encryption keys (SSE-C). Once enabled, new data is stored as encrypted with keys generated and kept by Storware Backup & Recovery. For performance improvements, we also recommend using AWS Direct Connect to access S3. Otherwise, backups would be sent over the Internet, which could result in poor performance.

Note: S3 has a limit of 5TB per object. This means that depending on the virtualization platform and backup format used by export/import mode you may have a limit of 5TB per VM (if it is Proxmox VMA or Citrix XVA image-based backup) or per VM disk (in most cases). Bigger files are currently not supported.

Permissions

Depending on the selected mode, you may have different permission sets. For a single bucket, you need to use the access keys of a user that has the ability to control objects within the bucket over the specific bucket - here is an example of IAM policy:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Sid": "Stmt1568968204280",
      "Action": [
        "s3:DeleteObject",
        "s3:DeleteObjectTagging",
        "s3:DeleteObjectVersion",
        "s3:DeleteObjectVersionTagging",
        "s3:GetBucketTagging",
        "s3:GetBucketVersioning",
        "s3:GetObject",
        "s3:GetObjectRetention",
        "s3:GetObjectTagging",
        "s3:GetObjectVersion",
        "s3:GetObjectVersionTagging",
        "s3:ListBucket",
        "s3:ListBucketVersions",
        "s3:PutObject",
        "s3:PutObjectTagging",
        "s3:PutObjectVersionTagging",
        "s3:RestoreObject"
      ],
      "Effect": "Allow",
      "Resource": "arn:aws:s3:::BACKUP_DESTINATION_BUCKET/*"
    }
  ]
}

You can also use a predefined role and create a user from the AWS console: https://docs.aws.amazon.com/IAM/latest/UserGuide/id_credentials_access-keys.html#Using_CreateAccessKey

Note: It is recommended to periodically rotate your access/secret keys. More information can be found here: https://aws.amazon.com/blogs/security/how-to-rotate-access-keys-for-iam-users/. After changing the key in AWS, remember to update it in Storware Backup & Recovery as well.

Bucket replication

Even though S3 is a highly available service, you may want to be prepared in case of a region failure. We recommend following this guide https://docs.aws.amazon.com/AmazonS3/latest/dev/replication.html to set up bucket replication so that your data is replicated to another region in a worst-case scenario. Remember to point Storware Backup & Recovery to the replicated bucket in case of a disaster.

Glacier/Deep Archive support

Starting from version 3.9 of vProtect now Storware Backup & Recovery is able to move older backups to a Glacier/Deep Archive storage tier. In the S3 backup provider settings, you need to enable the Move old versions to other storage class toggle and provide extended retention settings.

Keep in mind that Storware Backup & Recovery will try to restore it to S3 with an expiration set to 2 days. You'll notice that although the task is running, no progress is taking place as it is waiting for the object to be restored from Glacier to S3. This may take several hours as Glacier doesn't provide instant access to archival data. Once this part is completed, Storware Backup & Recovery will proceed with regular restore from a temporary S3 object.

Costs

When storing backups in S3, additional charges will occur for stored backups. Retention setting in Storware Backup & Recovery can limit the storage costs of stored backups.

Please visit https://aws.amazon.com/s3/pricing/ to check current AWS S3 pricing.

Example

Now we will show you how to quickly create S3 storage and integrate it with Storware Backup & Recovery as a backup destination. After logging in, expand the services tab a choose S3 under the Storage section:

Now create a new bucket for your backups:

In "Configure options" activate versioning: (In all other tabs, you can leave the default settings)

After creating a bucket, we need to create a new user with appropriate permissions:

Remember to choose the "Programmatic access" account type:

From the predefined roles, you can choose "AmazonS3FullAccess" or you can create a new one as described in the Permissions section:

Remember to download the .csv or copy the key credentials manually:

Now go to the Backup destination tab on the Storware Backup & Recovery dashboard and change the sub-tab to object storage. Provide the bucket name and key credentials, and then configure the remaining options according to your requirements: