The first major release of the open source Gluster distributed file system software since December 2009 takes aim at container integration and storage management improvements for hybrid and multi-cloud deployments.
The Gluster development community rewrote the management engine for the new Gluster 4.0 release with the goals of boosting scalability and making the open source storage software easier to use, according to Amye Scavarda, the Gluster community lead at Red Hat.
“The last time that we did this, around 3.0, was a long time ago, and the world was a lot different then. Gluster and containers were not even something we thought of together,” Scavarda said. “So, we’ve worked on being able to replace some of the architecture around it to be able to have a more complete service-oriented architecture.”
New GlusterD2 management engine
The GlusterD2 (GD2) distributed management engine, which is available as a technical preview, exposes functionality through a set of REST-based application programming interfaces (APIs). GD2 includes a new command line interface (CLI) on top of the REST API for volume and membership operations. It also includes a flexible plugin framework for developers to add metrics.
The Gluster 4.0 release also deepens connections with the Kubernetes container orchestration framework through its integration with the open source Heketi provisioning tool, which downloads with the Gluster file system (GlusterFS) software.
The latest version of Heketi supports provisioning and expanding Gluster-block backed persistent volumes, custom volume names for persistent volumes, and metrics collection for Gluster volumes using the separately downloadable open source Prometheus tool.
“Heketi is a dynamic provisioner. It’s basically a way to be able to say, ‘Hey, I want to be able to create a volume. I want to be able to specify the size, the replication factor, and I don’t want to have to think about how Gluster needs that. I want to be able to just give it what I need and have Heketi go in and work with Gluster for that directly,” she said.
Prior Gluster versions used Heketi, but the new Gluster 4.0 major release now connects it with the new GlusterD2 management engine, Scavarda said. She said the Gluster community has worked on improving container integration for the past three or four years, with a special focus on persistent storage and the integration with Kubernetes. She said Gluster users would now have a better way to expand volumes and replace disks and nodes.
Gluster focus on cloud use
“We’re really trying to think about the ways that people are using Gluster in the cloud,” Scavarda said. “They need to be able to actually use a system that already has storage and expand from there.”
Scavarda said that in addition to Kubernetes, Gluster also pairs well with Red Hat’s OpenShift container application platform. GlusterFS can be provisioned as the storage for a container environment, and the software can also run within a container.
New capabilities on the roadmap for future Gluster community releases include GlusterD2 management engine improvements and a graphical user interface, since many developers who write container-based applications are not storage experts, Scavarda confirmed.
Prominent users of the open source Gluster community software include cloud hosting and media companies, many of which like to test out the file system try the new features, according to Scavarda.
She could give no specific date when the commercially supported Red Hat Gluster Storage product would add the new Gluster 4.0 features. She said the capabilities in community releases tend to reach the supported product at least three to six month later.