Enhancements

With this release, you can now deploy NVME over Fabrics either by using the ceph orch apply nvmeof command or by applying a service specification with the service_type nvme.

(BZ#2239892)

With this enhancement, to safeguard data integrity, confidentiality, and alignment with the security best practices, TLS is enabled across the monitoring stack. The enhanced security feature for Prometheus, Alert manager and Node exporter adds an additional layer of protection by using secure communication across the monitoring stack.

(BZ#1994251)

With this enhancement, to safeguard data integrity, confidentiality, and to align with the security best practices, the authentication feature for Prometheus, Alert manager, and the Node exporter is implemented. This enhances the security of the whole monitoring stack by enabling TLS in all the monitoring components and requiring users to provide valid credentials before accessing Prometheus and Alert Manager data. By using this new feature, an additional layer of protection is provided by using secure communication and preventing the unauthorized access to sensitive metrics and monitoring data. With TLS enabled in all the monitoring stack components and authentication in place, users must authenticate before accessing monitoring and metrics data, enhancing overall security and control over data access.

(BZ#2028335)

Previously, users could not put a host into maintenance mode as stopping all the daemons on that host would cause data unavailability.

With this enhancement, a stronger force flag for the ceph orch host maintenance enter command called --yes-i-really-mean-it is added. Users can now put a host into maintenance mode, even when Cephadm warns against it.

(BZ#2107339)

With this enhancement, users can drain a host of daemons, without also draining the client conf or keyring files deployed on the host by passing the --keep-conf-keyring flag to the ceph orch host drain command. Users can now mark a host to have all daemons drained or not placed there while still having Cephadm manage conf or keyring files on the host.

(BZ#2153827)

Previously, the only way to mark the Cephadm services as managed or unmanaged was to edit and re-apply the service specification for the service. This was inconvenient for scenarios, such as temporarily stop making OSDs on devices that match an OSD specification in Cephadm.

With this enhancement, commands are added to set the Cephadm services to be marked as managed or unmanaged. For example, one can run ceph orch set-unmanaged mon command or ceph orch set-managed mon command. These new commands allow toggling the states without having to edit and re-apply the service specification.

(BZ#2155625)

Previously, users could not apply client IP restrictions, while still using HAProxy between the client and NFS. This is because only the HAProxy IP would be recognized by NFS, making proper client IP restriction impossible.

With this enhancement, it is possible to deploy an NFS service in HAProxy protocol mode by passing --ingress-mode=haproxy-protocol argument in the ceph nfs cluster create command or by setting enable_haproxy_protocol: true in both the NFS service specification and the corresponding ingress specification. Users can now apply proper client IP restriction on their NFS deployment using the new HAProxy protocol mode in their NFS deployment.

(BZ#2176300)

Previously, monitoring performance dumps from the MDS would sometimes show that the OSDs were laggy, objecter.op_laggy and objecter.osd_laggy, causing laggy clients and the dirty data for cap revokes would not be flushed.

With this enhancement, if the defer_client_eviction_on_laggy_osds parameter is set to true and a client gets laggy because of a laggy OSD then client eviction does not take place until OSDs are no longer laggy.

(BZ#2228066)

Previously, the quota.max_bytes value was set in bytes, resulting in often very large size values, which was hard to set or changed.

With this enhancement, the quota.max_bytes values can now be set with human-friendly values, such as M/Mi, G/Gi, or T/Ti. For example, 10GiB or 100K.

(BZ#2091154)

For more information, see Configuring a multi-site.

(BZ#2190355)

(BZ#2238472)

Previously, an IBM Storage Ceph cluster could only be upgraded through the command-line interface.

With this enhancement, you can easily view and upgrade the available versions of the storage cluster and also track the upgrade progress from the Ceph dashboard.

(BZ#2232295)

With this enhancement, an Overview page is added in the Object Gateway section of the Ceph dashboard. From this Overview page, users can now access a dedicated object gateway overview in the Ceph dashboard. This feature enhances the user experience by providing insights into the Object Gateway performance, storage usage, and configuration details. When multi-site is configured in the cluster, the user can also see the multi-site sync status directly on this Overview page.

(BZ#2233356)

With this enhancement, a new capacity usage progress bar is visible within the Block > Images table on the Ceph dashboard. The bar provides visible usage information, along with a percentage.

(BZ#2151171)

Previously, Ceph File System (CephFS) volumes could only be managed through the command-line interface.

With this enhancement, CephFS volumes can now be listed, created, edited, and removed through the Ceph dashboard.

For more information, see Configuring CephFS volumes.

(BZ#2232297)

Previously, Ceph File System (CephFS) subvolumes and subvolumes groups could only be managed through the command-line interface.

With this enhancement, CephFS subvolumes and subvolume groups can now be listed, created, edited, and removed through the Ceph dashboard.

For more information, see Configuring CephFS subvolumes and Configuring CephFS subvolume groups.

(BZ#2232298)

Previously, short hostnames were picked to resolve the Ceph Object gateway hostname which would cause issues.

With this enhancement, in the Ceph dashboard, the rgw_dns_name configuration option of the Ceph Object Gateway is used to resolve the hostname if it is provided. If you wants to specify the FQDN for the Ceph Object Gateway host, use the rgw_dns_name configuration option in the CLI and the dashboard then picks it up and the Ceph Object Gateway requests are made against it.

(BZ#2236109)

With this enhancement, S3 WORM feature is certified with an external entity. This enables data retention in compliance with FSI regulations and a secured object storage deployment that guarantees data retrieval even if the object/buckets in the production zones have been lost or compromised.

For more information, see Introduction to WORM.

(BZ#2068491)

With this enhancement, Ceph Object Gateway multisite sync instances dynamically parallelize workload among themselves, using a work sharing algorithm. As a result, the scalability is now significantly improved and sync workloads are divided evenly throughout the sync.

(BZ#1740782)

Bucket granular multi-sync sync replication was previously available as limited release. This enhancement provides full availability for new and existing customers in production environments. This feature allows enabling and disabling multi-site async replication on a per bucket level. This enhancement also increases the total RAW space required and increases the amount of synchronization traffic between sites.

For more information, see Bucket granular sync policies.

For more information, see S3 select content from an object.

This feature enables data transition in a Ceph Object Gateway storage environment, into the Azure cloud, for archiving purposes.

For more information, see Transitioning data to Azure cloud service.

For more information, see Integrating Ceph Object Gateway with Trino.

Previously, slow operations were marked as delayed but without a detailed description.

With this enhancement, you can view the detailed descriptions of delayed sub-events for operations.

(BZ#2240832)

Previously, there were no dedicated performance counters for accounting encrypted traffic in messenger v2.

With this enhancement, msgr_recv_encrypted_bytes and msgr_send_encrypted_bytes performance counters are introduced to account for receiving and sending bytes respectively that facilitate rough validation of the encryption status.

(BZ#1846154)

With this enhancement, HAProxy uses load balancing servers. This allows load balancing and also enables client restrictions on access.

(BZ#2097490)

With this enhancement, you can add the optional cmount_path option and generate a unique user ID for each Ceph File System to allow sharing CephFS clients across multiple Ganesha exports thereby reducing the memory usage for a single CephFS client.

(BZ#2236325)