Chapter 3. Enhancements


This section describes the major enhancements introduced in Red Hat OpenShift Data foundation 4.21.

3.1. Disaster Recovery

The Protected Applications page shows managed (ApplicationSet and Subscription) and discovered applications together, providing a complete view of all workloads protected by Disaster Recovery. Multi‑selection is enabled in the list view, laying the groundwork for grouped failover, relocate, and disable‑DR operations for managed applications.

Disaster recovery workflows now automatically remove resources associated with DR‑protected virtual machines during failover or relocate operations. This automation applies only to the discovered virtual machines enrolled for protection through Red Hat Advanced Cluster Management Fleet Virtualization and helps reduce manual steps while improving transition efficiency.

OpenShift Data Foundation now supports configuring Multus on existing clusters, enabling customers to introduce public and private storage networks after initial deployment. This enhancement allows administrators to add or expand Network Attachment Definitions (NADs), such as separating client‑facing and replication traffic even when a cluster is already using a single Multus network. This provides greater flexibility for network isolation, performance tuning, and aligning storage traffic with organizational requirements.

For more information, see Enabling Multus support on an existing OpenShift Data Foundation cluster.

Multiple device classes, Ceph pools, and StorageClasses can now be configured on the same disk type and shared nodes, enabling improved workload isolation, reduced noisy‑neighbor impact, and better support for multi‑tenant cost management. The enhancement also includes corrected OSD device‑class labels and UI updates to simplify configuration.

For more information, see Attaching storage for a new device set by using disks of same type on the same nodes.

3.4. Multicloud Object Gateway

Minimal IAM support is now available in the Multicloud Object Gateway to provide improved access control for multi‑tenant environments. This enhancement enables administrators to define and manage user‑level permissions more effectively, supporting scenarios where data scientists and other user groups require isolated and governed access to object storage resources. The feature lays the groundwork for more secure, flexible multi‑tenant deployments by ensuring that users interact only with the buckets and operations they are authorized to access.

For more information, see Creating and managing IAM user.

Multicloud Object Gateway (MCG) database backup is enhanced to provide more reliable protection against data loss and to complement the high‑availability capabilities. The update adds support for automated scheduled backups with configurable frequency (daily, weekly, or monthly) and retention settings along with on‑demand backup options through the MCG CLI or CRD. The UI now displays the time of the most recent backup and offers advanced configuration for Multicloud Object Gateway backup management. The system supports internal, external, and standalone MCG deployments by automatically detecting or allowing selection of snapshot StorageClasses.

To configure MCG database backup, refer to the deployment procedures for your specific platform.

Developer‑level users can now be granted access to the OpenShift Data Foundation Object Browser, enabling them to view and manage only their own object storage resources with restricted permissions. This enhancement supports flexible, role‑appropriate access control and allows development teams to work more independently while preserving administrative boundaries.

For more information, see Accessing the object browser.

Multicloud Object Gateway (MCG) performance has been improved by directing read‑only database operations to the existing secondary database instance, reducing load on the primary and improving overall responsiveness. By separating read and write activity, the system minimizes the impact of heavy read queries, such as lifecycle and background worker operations on write performance and object I/O. This enhancement leverages the high‑availability secondary instance already deployed in recent versions to deliver more consistent and efficient database behavior.

For more information, see the DB pods section within Responsibilities and resources.

The Object Browser now supports browsing, uploading, downloading, and managing RGW object. Operations include listing and creating buckets, navigating objects in a folder‑like view, uploading files or directories, downloading objects, generating presigned URLs, viewing metadata, filtering and sorting results, previewing content, and creating prefixes. Bucket‑level capabilities such as configuring expiration and access policies are also available.

For more information, see Creating and managing buckets using the object browser.

New Azure-based deployments now automatically enable Performance Plus for eligible Standard SSD and premium SSD disks (≥513 GiB) when using the Azure Disk CSI driver. This provides higher IOPS and throughput for OpenShift Data Foundation clusters on Azure and helps to achieve the best possible storage performance.

For more information about how to use, see the Prerequisites in the section Creating OpenShift Data Foundation cluster.

3.6. Monitoring enhancements

Additional NooBaa bucket and replication metrics

Introduces bucket_last_cycle_total_objects_num, bucket_last_cycle_replicated_objects_num, and bucket_last_cycle_error_objects_num, along with expanded bucket‑usage gauges for object and size quotas.

Clearer NooBaa capacity alerting

Adds concise runbooks for the NooBaaSystemCapacityWarning85, 95, and 100 alerts, outlining how capacity is calculated and the required mitigation steps.

For more information, see NooBaaSystemCapacityWarning85, NooBaaSystemCapacityWarning95, and NooBaaSystemCapacityWarning100.

New MDS xattr‑latency alert:

Adds the CephXattrSetLatency alert for elevated setxattr latency, with guidance for evaluating and adjusting MDS resource allocation.

For more information, see CephXattrSetLatency.

OpenShift Data Foundation now automatically detects IBM Z and LinuxONE architectures and applies optimized vCPU requirements specific to these platforms. Deployments on IBM Z and LinuxONE no longer require the same number of vCPUs as other architectures, reducing resource overhead and improving deployment efficiency.

Red Hat logoGithubredditYoutubeTwitter

Learn

Try, buy, & sell

Communities

About Red Hat Documentation

We help Red Hat users innovate and achieve their goals with our products and services with content they can trust. Explore our recent updates.

Making open source more inclusive

Red Hat is committed to replacing problematic language in our code, documentation, and web properties. For more details, see the Red Hat Blog.

About Red Hat

We deliver hardened solutions that make it easier for enterprises to work across platforms and environments, from the core datacenter to the network edge.

Theme

© 2026 Red Hat
Back to top