TL;DR

Optimal NFS mount options: rsize=1048576,wsize=1048576,nconnect=8,noatime,hard. Use NFSv3 for write destinations where timestamps matter. syncopio configures optimal mount options automatically — including NFSv3 fallback for timestamp-sensitive transfers.

These are the mount options we recommend for NFS data migrations. Copy this, tweak the IP and export path, and you’re done:

mount -t nfs -o rw,hard,noatime,rsize=1048576,wsize=1048576,nconnect=8 \
  192.168.1.100:/export/data /mnt/nfs

To make it persistent across reboots, add to /etc/fstab:

192.168.1.100:/export/data /mnt/nfs nfs rw,hard,noatime,rsize=1048576,wsize=1048576,nconnect=8 0 0

Option Reference

Option	What it does	Recommended
`rsize=1048576`	Read buffer size (bytes)	1 MB — max for NFSv3/v4
`wsize=1048576`	Write buffer size (bytes)	1 MB — matches rsize
`nconnect=8`	Parallel TCP connections per mount	4-16 — ~3x throughput on 10GbE
`noatime`	Skip access-time updates on reads	Always — reduces write I/O
`hard`	Retry indefinitely on server timeout	Yes — prevents silent corruption
`soft`	Return error after timeout	Avoid — risks incomplete writes
`intr`	Allow signal interrupts during hard hangs	Deprecated since kernel 2.6.25 — ignored on modern kernels
`nocto`	Skip close-to-open consistency checks	Only for read-only data
`actimeo=60`	Cache file attributes for 60 seconds	Good for mostly-static data
`vers=4.2`	Force NFSv4.2 (server-side copy, sparse files)	If server supports it

Version requirements

nconnect requires Linux kernel 5.3+. Check with uname -r before adding it — older kernels silently ignore the option.

Quick test

After mounting, run nfsstat -m to confirm your options took effect. Some servers silently downgrade nconnect or buffer sizes.

syncopio advantage

syncopio auto-mounts NFS endpoints with optimized options — no manual mount commands needed.

NFS Mount Options Cheat Sheet

Option Reference

Further Reading

Ready to simplify your migrations?