Longhorn, persistent storage software for Kubernetes clusters
Stemming from the development of Rancher Labs, publisher of a Kubernetes cluster management software stack, the Longhorn project covers the persistent storage of containerized applications. It is now hosted by the CNCF. Test for several months in beta version, it has just been delivered in general availability.
Sent by Rancher in October 2019 to the Cloud Native Computing Foundation, which integrated it as a sandbox project, the Longhorn cloud native persistent storage solution has been in general availability for a few days. It is designed for containerized applications managed with Kubernetes that need to store data persistently. These workloads have multiplied rapidly in production in recent months. The Longhorn project was initially started in 2017 by Rancher before joining the CNCF two years later.
While Rancher Labs ensures that it works closely with storage hardware vendors – NetApp, EMC and Pure Storage as well as Portworxs, StorageOS and OpenEBS – it also points to the existing demand for an open source storage solution that is less expensive and easier to deploy. What’s more, software that works straight from the start with its Rancher software stack that lets you deploy Kubernetes clusters in a datacenter, private or public cloud, bare metal or on vSphere VM. This stack centralizes authentication, access control and monitoring as part of global security policies. It also provides DevOps teams with built-in tools to run containerized workloads.
Cross-cluster disaster recovery with defined RTO and RPO
“With Longhorn, we took a fresh look at what ‘stateful’ workloads in Kubernetes require – primarily fast and reliable block storage,” said Tom Callway, Rancher Product Marketing Director, in a post. . The product has been in beta since last year and thousands of users have already tested its resistance, assures the publisher based in Cupertino (California). The version that ships this month into general availability includes a set of storage features for enterprises: fine allocation, snapshots, and backup / restore features. In addition, there are non-disruptive expansion and trans-cluster disaster recovery volume capacities with defined RTO (recovery time objective) and RPO (recovery point objective). The software is updated without impacting the operation of the volumes in operation. It integrates all of Kubernetes’ online commands and has its own user interface.
With Longhorn, we can set up distributed mirrored block storage on local disks. The solution is deployed on existing NFS, iSCSI and Fiber Channel storage arrays and on cloud storage such as AWS EBS providing application snapshots, backups and remote replication. One of Longhorn’s goals is to free itself from proprietary storage infrastructure and, in so doing, reduce costs. The solution itself is free to download. Rancher users will be able to install it from their catalog application. As for customers wishing to acquire premium support, they can do so as an add-on to their corporate subscription on Rancher.