Skip to main content

Plan for pNFS deployment

Contributors netapp-dbagwell

Before deploying pNFS in your environment, ensure you meet the prerequisites and understand the interoperability requirements and configuration limits.

Prerequisites

Before enabling and using pNFS in ONTAP, ensure the following requirements are met:

  • NFSv4.1 or later is enabled on the NFS server

  • At least one data LIF exists per node in the cluster for the SVM hosting the NFS server

  • All data LIFs in the SVM are routable to NFS clients

  • NFS clients support pNFS (most modern Linux distributions from 2014 and later)

  • Network connectivity between clients and all data LIFs in the SVM is functional

  • DNS resolution (if using hostnames) is configured properly for all data LIFs

  • FlexGroup volumes are configured (recommended for best results)

  • NFSv4.x ID domains match between clients and ONTAP

  • NFS Kerberos (if used) is enabled on all data LIFs in the SVM

Best practices summary

When implementing pNFS in your environment, follow these best practices:

Interoperability

pNFS in ONTAP is designed to work with RFC-compliant NFS clients. The following considerations apply:

Limits

The following limits apply to pNFS in ONTAP:

  • TCP connection limits per node vary by platform (check the NetApp Hardware Universe for specific limits)

  • Maximum file size: Depends on the volume type and ONTAP version

  • Maximum file count: Up to 200 billion files with FlexGroup volumes

  • Maximum capacity: Up to 60 PB with FlexGroup volumes

  • Network interface count: At least one data LIF per node is required; more might be needed for load balancing

When using nconnect with pNFS, be aware that TCP connection counts multiply quickly:

  • Each client mount with nconnect creates multiple TCP connections per data LIF

  • With many clients using high nconnect values, TCP connection limits can be exceeded

  • Exceeding TCP connection limits prevents new connections until existing connections are freed