CNCF welcomes Longhorn to its sandbox

CNCF welcomes Longhorn to its sandbox

Distributed block storage system for Kubernetes Longhorn has been accepted into the Cloud Native Computing Foundation’s sandbox.

The new home is meant to provide Longhorn with more public visibility, align it with other foundation projects, and ensure it follows the CNCF’s requirements which aim at low barriers for project adoption and contribution.

Longhorn was introduced to the microservice space by container tooling company Rancher Labs in April 2017. Back then, co-founder Sheng Liang described the project as a “new way to build distributed block storage for container and cloud deployment models”. It uses container orchestration to coordinate lightweight storage controllers to form a resilient distributed storage system.

According to its documentation, the projects stands – resilience aside – out for its simplicity, incremental snapshot capability and backup features. Non-disruptive upgrades and an user-friendly dashboard are also meant to convince users of the projects merits.

Longhorn doesn’t have to play on its own. Other projects currently populating the CNCF sandbox include monitoring tool Thanos, security projects Falco and in-toto, operators Flux, Strimzi, and KubeVirt, as well Virtual Kubelet, and KubeEdge.

The next step for the storage system would be the move into incubation status, which could happen after one of the annual reviews a project has to submit to. To get there, a project is required to have evidence for production use by at least three independent end users and a healthy number of committers. It also needs a clear versioning scheme and must demonstrate active development by a steady flow of commits and merged contributions.

Getting the necessary number of uses in production might not be that hard, since Rancher works closely with a variety of companies who, given enough support, might be open to install the project and be put on the list of references. It might, however, still take a while, because Longhorn is still in its alpha-phase and therefore not an intuitive choice for a dependable system.