The first release of Blockstor, an open distributed block storage management system for Kubernetes that provides data replication on top of DRBD, is now available. Blockstor is REST API compatible with LINSTOR and is able to work seamlessly with the existing ecosystem of clients, including the linstor command utility, CSI driver, Piraeus operator, ha-controller, and the golinstor library. The project is a completely independent (clean-room) implementation in the Go language, which does not use the original source code. The code is distributed under the Apache 2.0 license and developed within the Cozystack platform (CNCF Sandbox project).
The author of the project is Andrey Kvapil, founder of Cozystack and member of the Piraeus organization, within which the LINSTOR operator and CSI driver for Kubernetes are being developed. Kvapil is known in the Kubernetes community as a popularizer of LINSTOR and has given technical reports on the topic. Initially conceived as a small “Friday” initiative, the development of Blockstor turned into about 20 days of continuous work. The project is currently being developed as a research project but is considered a possible replacement for LINSTOR as the default storage system in Cozystack.
The reasons for creating a new project include difficulties in maintaining the original project, transferring changes to the main project, and architectural limitations of LINSTOR. Blockstor’s architecture is completely based on the Kubernetes controller-runtime approach. The configuration and current state of the system are presented in the form of Kubernetes CRD objects, and the system itself is not designed to work outside the Kubernetes cluster.
Some of the main features of Blockstor include volumes replicated on top of DRBD based on various backends, automatic placement of replicas considering zones and node properties, support for TieBreaker, quorum, and resizing volumes without stopping work, ability to work without DRBD in local or diskless storage mode, encryption of volumes via LUKS, support for snapshots, creation of storage pools from physical disks, and container images collected for different architectures published in GHCR.