Prevent NFS Mount delays by enforcing TCP

IN THIS ARTICLE

Outlines how to enforce TCP during NFS mounts to prevent delays

REQUIREMENTS

• Cluster running Qumulo Core

DETAILS

The use of UDP protocol during NFSv3 mount phase could cause a delay in the availability of the mount point since UDP is not supported by our product.

The above issue can potentially be exacerbated in environments where round-robin DNS is utilized due to the sequential mount attempts using UDP fail against each node in the cluster. This can result in a considerable lag until the NFSv3 export finally mounts using TCP.

We strongly encourage adding the TCP option to any NFS mount strings in auto-mounter maps to avoid this behavior using the example command below:

/mountpoint -fstype=nfs,nfsvers=3,tcp cluster_name:/path/dir

NOTE:  This article applies only to NFSv3, NFSv4.1 cannot use UDP. 

RESOLUTION

You should now be able to enforce TCP during NFS mounts to prevent delays.

ADDITIONAL RESOURCES

QQ CLI: Networks and IP Addresses

Required TCP ports for Qumulo Core

Like what you see? Share this article with your network!