0

I am using 12.04.4 and trying to connect to a NFS server which allows a blocksize of 64kbps, but the most I can set on the client side with mount.nfs is rsize/wsize=32786. I read that there is a kernel define NFSSVC_MAXBLKSIZE that may be limited to 32k, but I am not sure this is for client or server or both. Whichever, how can I change this, or something else to get a higher buffer size? Thanks,

2 Answers 2

0

I don't know a lot about this subject, but i made some research and i found that you can easily modify rsize and wsize values while mounting, The maximum value that can be set depends on the current kernel version you have.

Take a look at this : How to do Linux NFS Performance Tuning and Optimization

0

According to nfs(5), rsize and wsize can be set to 1048576 (1 MB), which is the maximum value for the NFS client. Older NFS versions (e.g. NFSv2) had smaller max values.

This can be changed as a mount option, e.g.:

sudo mount -t nfs -o rsize=1048576,wsize=1048576 server:/data /mnt/data

Note: the client and server negotiate the largest rsize/wsize value that they can both support. So, if the server doesn't support a rsize/wsize that large, a smaller one will be negotiated:

$ mount | grep /mnt/data | egrep -o rsize=[0-9]*
rsize=131072
3
  • 1
    Is there anyway to determine if the size is being set by negotiation? Is there anyway to enable this level of debug?
    – CptanPanic
    Commented May 22, 2015 at 11:04
  • "If" the size is set? The size is always negotiated, because both sides need to agree on a size. "Which" size is being set can be seen via "mount", see my answer above.
    – ckujau
    Commented May 22, 2015 at 19:46
  • So I found that if I use 'nfsstat -m' it shows the actual buffer size that is being used. And verified that the max allowed for me is 32k.
    – CptanPanic
    Commented May 26, 2015 at 11:41

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .