LinkedIn Sourceforge Twitter

Vincent's Blog

Pleasure in the job puts perfection in the work (Aristote)

NFS settings for a Linux client connected to an OpenBSD NFS server.

Posted on 2016-10-20 21:48:00 from Vincent in NFS

You are an happy owner of a NAS running on OpenBSD proposing NFS shared drives. And your life is perfect with OpenBSD client (maybe you have read one of my article on that subject). But this idilyc echosystem is in danger when a friend of you comes and connects is brand new Linux laptop on your NFS server.

In such case, often the default settings proposed by Linux distribution will not allow you to transfer files.

Let me share some mount parameters that should work much better:

  mount -t nfs -o rw,intr,soft,noatime,rsize=2048,wsize=2048,vers=3    <nfs-server-name>:/nfs/path /mounting/path

Some explanations:

  • -t nfs : do I have to explain it ?
  • -o : is the entry for optional parameters
  • rw : read and write
  • intr : interruptible
  • soft : as opposed to hard mount. "soft" is better for laptops
  • noatime : because I don't care to record the last access time on each file. This is a frequent tips to allow better performance
  • rsize and wsize : are packet size for read and write. Higher is better, but is not always a good idea to increase it too far.
  • vers : because Linux machines are using NFSv4, I force the usage of v3 (NFSv4 is not implemented in OpenBSD). I've tested with v2 and this works too.
  • then you have the NFS's server name or IP and his path.
  • the last item is the mount point.

I've tried in UDP and TCP. In each case I'm able to transfer big files at the optimum speed for my setup.

2, 2
displayed: 474

How much does 6 plus 2 ?