[Gluster-users] GCS 0.1 release!

[Atin Mukherjee]

== Overview

Today, we are announcing the availability of GCS (Gluster Container
Storage) 0.1. This initial release is designed to provide a platform for
community members to try out and provide feedback on the new Gluster
container storage stack. This new stack is a collaboration across a number
of repositories, currently including the main GCS repository [1], core
glusterfs [2], glusterd2 [3], and gluster-csi-driver [4].

== Getting started

The GCS repository provides a VM-based (Vagrant) environment that makes it
easy to install and take GCS for a test-drive. See
https://github.com/gluster/gcs/tree/master/deploy#local-cluster-using-vagrant
for a set of instructions to bring up a multi-node cluster with GCS
installed. The Ansible-based deploy scripts create a Kubernetes cluster
using kubespray, then deploy the GCS components. These playbooks can also
be used to bring up GCS on other Kubernetes clusters as well.

== Current features

This is the initial release of the GCS stack. It allows dynamic
provisioning of persistent volumes using the CSI interface. Supported
features include:

-

1x3 (3-way replicated) volumes
-

GCS should be able to recover from restarts of any individual
GCS-related pod. Since this is the initial version, bugs or feedback on
improvements will be appreciated in a form of github issue in the
respective repos.


== Next steps

-

Adding e2e testing for nightly validation of entire system
-

Will be adding gluster-prometheus for metrics. The work under this can
be tracked at the gluster-prometheus repo [5]
-

Starting work on operator to deploy and manage the stack through anthill
[6]
-

Bi-weekly update to the community on the progress made on GCS.


== GCS project management

-

GCS and the other associated repos are coordinated via waffle.io for
planning and tracking deliverables over sprints.


-

Cross-repo coordination of milestones and sprints will be tracked
through a common set of labels, prefixed with “GCS/”. For example, we
already have labels for major milestones defined like ‘GCS/alpha1’ ,
‘GCS/beta0’. Additional labels like 'GCS/0.2'/'GCS/0.3'/... will be created
for each sprints/releases so that the respective teams can tag planned
deliverables in a common way.


== Collaboration opportunities

-

Improving install experience
-

Helping w/ E2E testing framework
-

Testing and opening bug reports


== Relationship to Heketi and glusterd (the legacy stack)

While GCS is shaping the future stack for Gluster in containers, the
traditional method for deploying container-based storage with Gluster (and
current GlusterD) and Heketi is still available, and it remains the
preferred method for production usage. To find out more about Heketi and
this production-ready stack, visit the gluster-kubernetes repo [7].

Regards,

Team GCS

[1] https://github.com/gluster/gcs

[2] https://github.com/gluster/glusterfs

[3] https://github.com/gluster/glusterd2

[4] https://github.com/gluster/gluster-csi-driver/

[5] https://github.com/gluster/gluster-prometheus

[6] https://github.com/gluster/anthill

[7] https://github.com/gluster/gluster-kubernetes