I'm looking for maximum of performance for a NFS mount.

I'm on local network, it could be completely unsecured, the performance is the most important in my case.

What I already found :

In /etc/export I've defined async mode, and it already helped a lot.

In /etc/default/nfs-kernel-server I've found following variables, which could potentially increase the performance :

# Number of servers to start up
# Runtime priority of server (see nice(1))
# Options for rpc.mountd.

The NFS server is on Linux Mint, and the client, which mount remote partition is Debian 6.

If there are any recommendation for maximum performance of NFS server ?

Should I change the RPCNFSDPRIORITY ?

If there are any other parameters, which could boost NFS server ?

Client actually mounts the NFS partition with following configuration :

server:/remotepath    /localpath    nfs4 rw,intr,hard,noatime,rsize=32768,wsize=32768 0 0

Is that optimal mounting options for NFS v4 share ?

  • You mentioned you don't care if it's unsecure, then try to go with nfs v3. but if you find something more concrete I would love to know as well. – BitsOfNix Mar 27 '13 at 13:43
  • If I understand good, v4 is more performant, than v3, at least it's mentionned at en.wikipedia.org/wiki/Network_File_System – Fedir RYKHTIK Mar 27 '13 at 14:32

You can modify timeo and retrans options while mounting. For example:

mount /mnt -o timeo=5,retrans=4 

this affects the number of retry attempts made by the client to the server in case of a delayed response from the server or sometimes no response from the server.

| improve this answer | |

Don't hardcode rsize and wsize on modern Linux, they are automagically optimized and will get bigger than your example.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.