Lustre share
Native Lustre mounts
Point a PV at a pre-provisioned Lustre share and KlustreFS makes sure pods mount it using familiar mount.lustre options.
What is KlustreFS?
KlustreFS is a CSI driver that mounts your existing Lustre shares straight into Kubernetes, bridging the gap between HPC and cloud-native infrastructure. Use the same PV/PVC workflow you already know while letting AI training workloads and other persistent pods access high-throughput (RWM) scratch space—no storage re-platforming required.
Why KlustreFS
Everything you need to bring Lustre into Kubernetes
HPC job
Portable HPC pipelines
Reuse the same Lustre-backed scratch space whether the job runs under SLURM or Kubernetes; PVCs keep the paths you already trust.
Namespaced PVC
Straightforward PVC bindings
Keep each PVC pointed at a single Lustre export so pods only mount the directories Lustre admins already own.
Static PV definitions map directly to Lustre exports.
Uses the Lustre client already installed on worker nodes.
ReadWriteMany shared scratch volumes for pods.
Keep storage flows identical when moving jobs off SLURM.