v1.2.0

Added

• A new BeeGFS CSI Driver Operator as an option to deploy and manage the lifecycle of the driver. This allows for a more seamless discovery and installation experience from clusters running Operator Lifecycle Manager (OLM).
• Documentation and job specifications showing how to deploy the driver to HashiCorp Nomad.
  • Note: At this time the BeeGFS CSI driver does not officially support Nomad. These are being provided as an example for others who might want to experiment with using BeeGFS and Nomad, in particular anyone interested in contributing to any future efforts around Nomad.
• Documentation on how to deploy the driver to Kubernetes clusters where some nodes can access BeeGFS volumes, and some cannot.
• Support for BeeGFS v7.2.4, Kubernetes v1.21, and RedHat OpenShift v4.8.
• Support for specifying BeeGFS mount options on a persistent volume or storage class.
• Information on how to contribute to the project.

Changed

• Greatly improved performance of end-to-end testing by parallelizing many tests and being more selective about when certain tests run.
• Updated the project to adhere to v1.5.0 of the CSI specification.

Fixed

• Automated tests failing in a confusing manner when csi-beegfs-config.yaml is empty.

Known Issues

• In some instances Kubernetes has been observed to call DeleteVolume prior to NodeUnpublishVolume and NodeUnstageVolume. This has the effect of leaving behind BeeGFS mount points on Kubernetes nodes for volumes that no longer exist in the Kubernetes API or BeeGFS. Over time if enough "orphaned" mounts accrue, the Kubernetes node may become unstable. To date this has only been observed as part of end-to-end testing, and is suspected to be either a side effect of how the E2E test framework interacts with Kubernetes, or a bug within Kubernetes itself.