Chapter 3. New features

This section lists all major updates, enhancements, and new features introduced in this release of Red Hat Ceph Storage.

3.1. The Cephadm utility

cephadm automatically updates the dashboard Grafana password if it is set in the Grafana service spec

Previously, users would have to manually set the Grafana password after applying the specification.

With this enhancement, if initial_admin_password is set in an applied Grafana specification, cephadm automatically updates the dashboard Grafana password, which is equivalent to running ceph dashboard set-grafana-api-password command, to streamline the process of fully setting up Grafana. Users no longer have to manually set the dashboard Grafana password after applying a specification that includes the password.

OSDs automatically update their Ceph configuration files with the new mon locations

With this enhancement, whenever a monmap change is detected, cephadm automatically updates the Ceph configuration files for each OSD with the new mon locations.

Note

This enhancement may take some time to update on all OSDs if you have a lot of OSDs.

3.2. Ceph Dashboard

The Block Device images table is paginated

With this enhancement, the Block Device images table is paginated to use with 10000+ image storage clusters as retrieving information for a block device image is expensive.

Newly added cross_origin_url option allows cross origin resource sharing

Previously, IBM developers faced issues with their storage insights product when they tried to ping the REST API using their front-end because of the tight Cross Origin Resource Sharing (CORS) policies set up in Red Hat’s REST API.

With this enhancement, CORS is allowed by adding the cross_origin_url option that can be set to a particular URL - ceph config set mgr mgr/dashboard/cross_origin_url localhost and the REST API allows communication with only that URL.

3.3. Ceph File System

Users can store arbitrary metadata of CephFS subvolume snapshots

With this enhancement, Ceph File System (CephFS) volume users can store arbitrary metadata in the form of key-value pairs for CephFS subvolume snapshots with a set of command-line interface (CLI) commands.

3.4. Ceph Object Gateway

STS max_session_duration for a role can now be updated

With this enhancement, the STS max_session_duration for a role can be updated using the radosgw-admin command-line interface.

ListBucket S3 operation now generates JSON output

With this enhancement, on customers’ request to facilitate integrations, the ListBucket S3 operation generates JSON-formatted output, instead of the default XML, if the request contains an Accept: application/json header.

The option to enable TCP keepalive managed by libcurl is added

With this enhancement, the option to enable TCP keepalive on the HTTP client sockets managed by libcurl is added to make sync and other operations initiated by Ceph Object Gateway more resilient to network instability. This does not apply to connections received by the HTTP frontend, but only to HTTP requests sent by the Ceph Object Gateway, such as Keystone for authentication, sync requests from multi-site, and requests to key management servers for SSE.

Result code 2002 of radosgw-admin commands is explicitly translated to 2

Previously, a change in the S3 error translation of internal NoSuchBucket result inadvertently changed the error code from the radosgw-admin bucket stats command, causing the programs checking the shell result code of those radosgw-admin commands to see a different result code.

With this enhancement, the result code 2002 is explicitly translated to 2 and users can see the original behaviour.

You can now use use bucket policies with useful errors

Bucket policies were difficult to use since the error indication was wrong. Additionally, silently dropping principals would cause problems during the upgrade. With this update, useful errors from policy parser and a flag to reject invalid principals with rgw policy reject invalid principals=true parameter is introduced.

3.5. Multi-site Ceph Object Gateway

The bucket sync run command provides more details

With this enhancement, user-friendly progress reports on the bucket sync run command are added to provide users easier visibility into the progress of the operation. When the user runs the radosgw-admin bucket sync run command with --extra-info flag, users get a message for the start of generation sync and also for each object that is synced.

Warning

It is not recommended to use the bucket sync run command without contacting Red Hat support.

Multi-site configuration supports dynamic bucket index resharding

Previously, only manual resharding of the buckets for multi-site configurations was supported.

With this enhancement, dynamic bucket resharding is supported in multi-site configurations. Once the storage clusters are upgraded, enable the resharding feature, zone level, and zone group. You can either manually reshard the buckets with radogw-admin bucket reshard command or automatically reshard them with dynamic resharding, independently of other zones in the storage cluster.

Users can now reshard bucket index dynamically with multi-site archive zones

With this enhancement, multi-site archive zone bucket index can be resharded dynamically when dynamic resharding is enabled for that zone.

3.6. RADOS

Low-level log messages are introduced to warn user about hitting throttle limits

Previously, there was a lack of low-level logging indication that throttle limits were hit, causing these occurrences to incorrectly have the appearance of a networking issue.

With this enhancement, the introduction of low-level log messages makes it much clearer that the throttle limits are hit.

3.7. RADOS Block Devices (RBD)

Cloned images can now be encrypted with their own encryption format and passphrase

With this enhancement, layered client-side encryption is now supported that enables each cloned image to be encrypted with its own encryption format and passphrase, potentially different from that of the parent image. The efficient copy-on-write semantics intrinsic to unformatted regular cloned images are retained.