Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News MayaData Launches Kubernetes Cloud-Native Storage Management Service Kubera

MayaData Launches Kubernetes Cloud-Native Storage Management Service Kubera

This item in japanese

MayaData announced the launch of Kubera - a management service for stateful workloads on Kubernetes that use OpenEBS as a storage layer.

Kubera is a management solution for stateful workloads for software like Kafka and Cassandra, which require persistent storage and use OpenEBS as a storage layer. Kubera also includes backups, upgrades, logging, analytics, and monitoring of OpenEBS workloads. OpenEBS is open source, and Kubera aims to ease operations around OpenEBS.

OpenEBS is a CNCF project for container-attached storage (CAS) on Kubernetes. It joined the CNCF in 2017 and was accepted as a sandbox project in 2019. CAS is a general term for software that "includes microservice based storage controllers that are orchestrated by Kubernetes". OpenEBS runs in user space - thus avoiding any Linux kernel module dependencies - and can be installed on a Kubernetes cluster either via Helm or using a Kubernetes manifest. In contrast to non-CAS systems, the storage controller as well as the storage replicas in a CAS system run as Kubernetes pods. Another difference between CAS and distributed storage is that "the metadata of volume replicas are not shared among the nodes and is independently managed on every local node".

MayaData's CEO Evan Powell explains this in a recent Kubernetes podcast:

What you're actually doing with OpenEBS is you're writing data to a number of replicas. Those replicas themselves are containerized. And that can be one replica which means you have one copy of your data. It can be three. It can be more.

Powell adds that the writes can be either synchronous or asynchronous. In the former case the client would not receive a success acknowledgement from the storage system until the data has been written to all replicas. The latter scenario makes more sense when some of the replicas might be on a different cloud, or the write itself might take a long time.

CAS also mitigates the "IO Blender Effect", which causes degraded storage performance for applications running in virtualized environments due to the hypervisor mixing input/output streams from different VMs. OpenEBS by itself supports replication, copy-on-write snapshots and backups and provides block storage for containers. Three types of backend storage are supported - Jiva, cStor, and local persistent volumes.

Kubera's features include automation of common workflows - upgrades and high availability, and operational features like cluster logging, analytics, visualization and monitoring. It also supports backups of OpenEBS as well as other workloads. Kubera can be connected to a Kubernetes cluster running on a public cloud, private cloud or an on-premise cluster.

OpenEBS has out of the box support for Prometheus for monitoring. Similar tools in this space are Portworx - which supports non-Kubernetes orchestrators also, Longhorn from Rancher Labs, and StorageOS. There are also some other companies which provide commercial services related to OpenEBS.

The source code for OpenEBS is available on GitHub. Kubera can be deployed on-premises or accessed as a SaaS platform.

Rate this Article