Chapter 2. New features
This section describes new features introduced in Red Hat OpenShift Data Foundation 4.18.
2.1. Disaster recovery solution
2.1.1. Awareness of replication delays ahead of failover or relocation
Non-blocking warnings are displayed for synchronization delays during failover or relocation operations for both discovered and managed applications. This helps to be aware of the replication delays due to any replication issues and replication delay until the initial sync is complete.
For more information, see Subscription-based application failover between managed clusters.
2.1.2. More recipe capabilities for RBD-based applications
The capabilities of recipes are enhanced to support many more RBD-based applications.
2.2. Scaling storage using multiple device classes in the same cluster for local storage deployments
Red Hat OpenShift Data Foundation supports the use and segregation of different disk types. These disk types can be segregated into different device classes and exposed as separate storage classes.
This way, it is possible to control which workloads receive which local storage with their specific disk performance.
Red Hat OpenShift Data Foundation only supports flash disks.
Different sets of disks can be used in the same cluster (local storage), which provides the flexibility to use multiple device classes in the same cluster.
For more information, see Scaling storage using multiple device class in the same cluster for local storage deployments.
2.3. Reduction in the time required to upgrade Red Hat OpenShift Data Foundation
Optimization done to reduce the time required to upgrade Red Hat OpenShift Data Foundation in OpenShift clusters with a significant number of nodes. The optimization takes into account the number of nodes and cluster configuration to better behave during upgrade. This helps to upgrade more than one CSI RBD or CephFS plugin pod at a time during the upgrade.
For more information, see the Prerequisites section in Updating Red Hat OpenShift Data Foundation 4.17 to 4.18.
2.4. Versioning for Multicloud Object Gateway Bucket Replication
Versioning in Multicloud Object Gateway (MCG) enables object data collaboration between two different locations or replication of object data from NooBaa on-prem to NooBaa on AWS, NooBaa to NooBaa, on any platform. This allows opting in for synchronizing versioning.
For more information, see Synchronizing versions in Multicloud Object Gateway bucket replication.
2.5. Bucket notification in Multicloud Object Gateway
Bucket notifications allow for a performant data pipeline creation. With this, it becomes easy to create a data flow where newly ingested data can be immediately detected and further processed. This is specifically important in AI/ML use cases.
For more information, see Bucket notification in Multicloud Object Gateway.
2.6. Multicloud Object Gateway object browser
Bucket content can quickly be browsed, uploaded and downloaded using the MCG object browser inside the OpenShift console. This provides a simplified way to browse the MCG object storage and avoids using a third party tool.
For more information, see Creating and managing buckets using MCG object browser.
2.7. Support for RADOS namespace for external mode
OpenShift Data Foundation supports RADOS namespace for its external mode clusters. This helps to improve performance in multi-tenants scenarios. By creating RADOS namespaces with restricted access per tenant on cephBlockPools will provide an efficient way of serving RBD storage.
For more information, see Creating an OpenShift Data Foundation Cluster for external Ceph storage system.
2.8. Ceph commands from the CLI tool
Arbitrary ceph CLI commands can be run as part of the ODF CLI tool, mainly for guided troubleshooting using the Red Hat OpenShift Data Foundation documentation.
Generally, configuration changes using ceph commands directly are not supported without explicit instructions from Red Hat support.