Fusion Backup & Restore - Backup Storage Location
When configuring Fusion Backup & Restore, you are required to create a backup location. When you create a backup location, you are given the following choices:
- Azure (Microsoft • Object Storage)
- IBM Cloud (IBM • Object Storage)
- AWS (Amazon • Object Storage)
- S3 Compliant (Any • Object Storage)
- Storage Protect (IBM • Object Storage and Tape)
What does "S3 Compliant" and "Any" exactly mean? What about other object storage functions, are those supported? The purpose of this document is to provide answers to frequently asked questions about Fusion Backup & Restore support of object storage.
Fusion Backup & Restore - Object Storage FAQ
Q: What does "S3 Compliant" and "Any" mean in terms of supported object storage?
A: S3 compliant (or S3 compatible) means that IBM Fusion uses Amazon's Simple Storage Services (S3) protocols and API specifications to read and write backup data to object storage. Any object storage system that claims to be "S3 compliant" or "S3 compatible" should be able to function with IBM Fusion Backup & Restore.
The exception to this general rule is that there are object storage systems that are not supported by OADP and therefore not supported by IBM Fusion Backup & Restore. Refer to Red Hat Documentation for a list of unsupported backup storage providers. Note that OADP does not currently support Google Cloud (even though it is not currently listed in their documentation).
Q: Does object storage for Fusion Backup & Restore have to be in the cloud?
A: No! Object storage can be in the cloud or on-premises. Examples of supported on-premises object storage that can be used with Fusion Backup & Restore: IBM Spectrum Protect, IBM Data Protect, OpenShift Multicloud Object Gateway to name a few.
Q: Does Fusion Backup & Restore support the Amazon S3 Glacier service?
A: No. The S3 Glacier specification is different then the general S3 specification and there is currently no support for this feature. Generally speaking, Glacier manages and stores data on low-cost storage (read "tape") and the S3 Glacier specification allows for a program to request data being staged back to disk before being able to read the data.
Q: Does Fusion Backup & Restore support S3 Object Lock?
A: No. Object lock is not supported at this time. If your target bucket supports object lock, Fusion Backup & Restore will be able to store data into the bucket but will not be able to delete data when data is expired. For example, if your Fusion backup policy is set to 30 day retention but your object storage bucket has an object lock for 728 days (2 years), after 30 days you will still be able to create backups in the bucket but you will not be able to delete expired data and therefore store more data then is intended by your 30 day retention policy.
Q: Are there any best practices when using object storage with IBM Fusion Backup & Restore?
A: If you are using on-premises object storage, you should make sure that the object storage is not hosted on the same system that you are trying to protect. Also consider using a separate bucket for OpenShift application/vms and for Fusion Service protection backups.
Q: Does Fusion Backup & Restore support object storage encryption?
A: The question needs to be answered in two parts:
- Encryption in-flight. Fusion Backup & Restore uses the https protocol to encrypt data between the OCP cluster and object storage target.
- Encryption at-rest. Fusion Backup & Restore does not advertise the encryption of data at rest. If the user requires at-rest encryption, they need to turn on encryption in their object storage service.
Q: Does Fusion Backup & Restore support sending data over the HTTP protocol?
A: Yes. Fusion Backup & Restore supports both HTTPS and HTTP protocols. You should use caution when using the HTTP protocol as data is not being sent securely to the object storage target. It is recommended that the HTTPS protocol is used in production environments and the use of the HTTP protocol is limited to testing purposes only.