Limitations in DataCore Puls8 v4.6

This section outlines the supported boundaries, constraints, and operational conditions that restrict or influence the behavior of certain features in this release.

The items below specify the current limitations:

NFS RWX Volumes Require a Replicated PV Mayastor Backend

The DataCore Puls8 NFS driver supports only Replicated PV Mayastor (io.openebs.csi-mayastor) as the backend for NFS RWX volumes. The backend StorageClass provisioner is validated at provisioning time, the bound PersistentVolume's CSI driver is validated on adoption, and the VolumeSnapshotClass driver is validated on snapshot. Any other backend is rejected.

Recommended Action: Set backendStorageClass in your NFS StorageClass to a Replicated PV Mayastor StorageClass.

NFS RWX Volume Feature Constraints

NFS RWX volumes are subject to the following constraints in this release:

  • Filesystem mode only. Raw block volumes are not supported.
  • Volume cloning is not supported. Create a snapshot and restore from it instead.
  • Per-volume usage statistics are not reported to the kubelet, so NFS RWX PVCs do not appear in kubelet_volume_stats_* metrics.
  • Each volume is served by a single NFS server pod. Availability during node failure is provided by the Lifeline Controller rather than by a clustered NFS configuration.
  • The exported path is fixed when the volume is created and cannot be changed afterwards.

V2 Volumes Consume Additional Pool Capacity for Metadata

Replicated PV Mayastor volumes created with the V2 label layout consume more pool space per replica than the volume's requested size. For example, a 10 MiB volume presents a 10 MiB device while each replica consumes 20 MiB. Volumes created before this release remain on the V1 layout and are unaffected.

Recommended Action: Allow additional headroom when sizing DiskPools, particularly on clusters hosting a large number of small volumes. Refer to Creating a DiskPool.

High Availability Not Supported on RHEL 10

High Availability (HA) functionality is not supported on RHEL 10 and its derivative distributions. This limitation is due to the removal of the nvme_core.multipath kernel parameter in RHEL 10, which is required for enabling volume failover in HA configurations. If you deploy DataCore Puls8 on RHEL 10-based environments, HA will not function as expected. Volume failover may fail, potentially leading to reduced availability and resilience.

Recommended Action: To use the HA feature, ensure you are running on a Linux distribution that supports the nvme_core.multipath kernel parameter.