Skip to content

ROADMAP updates

Jump to bottom
Kiran Mova edited this page Jan 18, 2022 · 1 revision

Roadmap

This document provides information on OpenEBS development in current and upcoming releases. Community and contributor involvement is vital for successfully implementing all desired items for each release. We hope that the items listed below will inspire further engagement from the community to keep OpenEBS progressing and shipping exciting and valuable features.

OpenEBS follows a lean project management approach by splitting the development items into current, near term and future categories.

Current

These are some of the backlogs that are prioritized and planned to be completed within the next major release (e.g. OpenEBS 3.0). While the following are planned items, higher priority is given to usability and stability issues reported by the community. The completion of these items also depends on the availability of contributors.

Note: OpenEBS follows a monthly release cadence with a new minor release on the 15th of every month. For the most current plan and status check out the release project trackers or the component specific trackers listed below. This document is reviewed and updated by the maintainers after each release.

Dynamic Local PVs

Mayastor

  • Source repositories
  • Backlogs
    • [Done] User applications can continue to access volumes when the nexus hosting them fails (e.g. Mayastor container crashes or is otherwise rescheduled, or its host node is lost or disconnected)
    • [Done] It should be possible for Moac (and all other significant control plane components) to be rescheduled within a cluster
    • [Done] Refactoring for better control plane and stability fixes
    • [In-progress] Mayastor Replica placement should be topology aware for statefulsets (zonal (or HA) distribution.
    • [In-progress] Mayastor should expose metrics that meet the needs of the SRE persona, to trend review throughput, latency, capacity utilization, and errors
    • [In-progress] Support for VolumeSnapshot
    • [In-progress] API refactoring exposed via gRPC
    • [In-progress] Allow a new replica to be created within the same Mayastor Pool as the failed replica it replaces
    • [In-progress] Failed replicas should be garbage collected (return capacity to Mayastor Pool)

Jiva

  • Source repositories
  • Backlogs
    • [Done] Enhance Jiva Operator functionality to reduce manual steps around launching new replicas when node is completely removed from the cluster
    • [Done] Add additional integration tests to Jiva CSI Driver to move towards beta
    • [Done] Consolidate the CSI driver and Jiva control plane into single repo
    • [Done] Documented the steps to migrate volumes from out-of-tree provisioners to CSI Driver
    • [In-progress] Deprecate the v1alpha1 CRDs in favor of the v1 CRDs introduced from 3.1

cStor

NDM

  • Source repositories
  • Backlogs
    • [Done] Enhance the discovery probes to identify virtual storage (without WWN) moving across nodes
    • [Done] Add gRPC API to list and re-scan block device
    • [Done] Enhance the discovery probes to detect if the device already has device mapper, zfs and so forth
    • [Done] Scan for device media errors and report them as prometheus metrics via ndm-exporter
    • [Done] Label the block devices so that they can be reserved for use by different StorageClasses
    • [Done] Auto-detecting capacity and mountpoint changes and updating the block device CR
    • [Done] Support for using a custom node label to claim devices (instead of default kubernetes.io/hostname)

Others

Near Term

Typically the items under this category fall under next major release (after the current. e.g 4.0). At a high level, the focus is towards moving the beta engines towards stable by adding more automated e2e tests and updating the corresponding user and contributor documents. To name a few backlogs (not in any particular order) on the near-term radar, where we are looking for additional help:

  • Support for Mayastor Volume resize
  • Support for pluggable storage backend for Mayastor (example: replace blobstore with lvm)
  • Support for specifying multiple hostpaths to be used with Local PV hostpath
  • Update user documentation with reference stacks of running various workloads using OpenEBS volumes
  • Auto provisioning of block devices (on the external storage systems) that can be used with OpenEBS storage engines
  • Setup E2e pipelines for ARM Clusters
  • Automate the workflows around handling scenarios like complete cluster failures that currently require some manual steps
  • Custom Kubernetes storage schedulers to address auto-rebalancing of the data placed on the nodes to help with scale up/down of Kubernetes nodes
  • Allow Mayastor Pools to incorporate more than one capacity contributing disk device
  • Auto-scaling up and down of cStor pools as the new nodes are added and removed
  • Auto-upgrade of cStor Pools and Volumes when user upgrades control plane
  • Asynchronous or DR replica for cStor and Mayastor volumes
  • Support for restoring a volume (in-place) for supporting blue/green stateful deployments
  • Multi-arch builds for all Mayastor components
  • Partial rebuild for the Mayastor replicas (similar to zfs resilvering)
  • Support Bulk BDC requests to claim multiple block devices that satisfy affinity or anti-affinity rules of applications. Example: two block devices from same node or two block devices from different nodes.
  • Support for device configuration tasks like partitioning, mounting or unmounting devices by adding new services via NDM gRPC API layer.

Future

As the name suggests this bucket contains items that are planned for future. Sometimes the items are related to adapting to the changes coming in the Kubernetes repo or other related projects. Github milestone called future backlog is used to track these requests.

Getting involved with Contributions

We are always looking for more contributions. If you see anything above that you would love to work on, we welcome you to become a contributor and maintainer of the areas that you love. You can get started by commenting on the related issue or by creating a new issue. Also you can reach out to us by:

Clone this wiki locally