Potentially smoothing the challenges that Kubernetes users have faced with stateful workloads, cloud native software and support provider Rancher Labs has donated its open source Longhorn Kubernetes block storage software to the Cloud Native Computing Foundation, where it will join 20 other current projects at CNCF’s Sandbox level.
“I think it’s a win-win. it’s a great win for the community because Longhorn is just fantastic technology and it’s the kind of technology that CNCF needs to further popularize technologies like Kubernetes. So, that’s great for the CNCF,” said Rancher Labs CEO Sheng Liang in an interview with The New Stack. “It’s also good for Rancher, because we have our hands full working on K3S [lightweight Kubernetes distribution] and Rancher. Rancher is not going to be a storage company. We have our hands full just monetizing and building the enterprise Kubernetes platform itself. By making it a part of the CNCF, we can attract a lot more people to use it.”
Gluser and Ceph were “designed to be run by some storage admin. In the Kubernetes world, a lot of these things tend to be deployed by DevOps teams, so [the storage layer] needs to be a lot more lightweight and a lot simpler.” — Rancher Labs CEO Sheng Liang.
Sandbox-level CNCF projects are the earliest of three levels and are intended to progress to incubation and finally graduation levels. As a sandbox project, the CNCF’s goals are to increase the project’s visibility, facilitate alignment with other projects, nurture the project via CNCF Service Desk requests, and to remove possible legal and governance obstacles to adoption and contribution.
“By becoming a CNCF Sandbox project, Longhorn can help the open source community accelerate the maturity of persistent block storage solutions for Kubernetes,” said CNCF Technical Oversight Committee member Liz Rice in a statement.
While Liang acknowledged that there is other storage software out there for Kubernetes providing persistent storage, both open source and proprietary, he said that a key distinction for Longhorn was ease-of-use and simplicity. Longhorn is written in just 30,000 lines of Go and is built on existing Linux storage technologies, which he said not only makes it more accessible to end-users, but also to developers who want to contribute to the project itself.
“Before Longhorn was ready, we were actually using Gluster a little bit and it didn’t quite suit our needs. A lot of people in the community also tried to use Ceph. These are very good and very, very mature, and very powerful open source storage systems,” said Liang. “If there’s one thing that’s slightly unsatisfactory, in my opinion, it’s that they are a little bit complex. It’s kind of designed to be run by some storage admin. In the Kubernetes world, a lot of these things tend to be deployed by DevOps teams, so it needs to be a lot more lightweight and a lot simpler.”
Longhorn, which is currently at version 0.6.2, doesn’t simply add persistent storage to Kubernetes-based applications, it also provides “volume snapshots, built-in backup and restore, live upgrades without impacting running volumes, cross-cluster disaster recovery with defined [Recovery Time Objective] and [Recovery Point Objective], one-click installation, and an intuitive user interface,” according to a company statement.
Among the highlighted features said to make Longhorn easier to use than other solutions is this user interface, which “exposes a storage class for easy provisioning of replicated volumes” and offers “visibility into storage volumes, but it also helps teams understand the Kubernetes workload that created the volume,” according to a company statement. Furthermore, Rancher notes that “Longhorn can be installed and upgraded with a few clicks, without needing to first read all of the documentation to understand every nuance.”
Longhorn is available on GitHub under the Apache 2.0 license.
The Cloud Native Computing Foundation is a sponsor of The New Stack.