> -----Original Message----- > From: Chuck Lever [mailto:chuck.lever@xxxxxxxxxx] > Sent: Wednesday, September 01, 2010 9:16 PM > To: Chuck Lever > Cc: Tayade, Nilesh; linux-nfs@xxxxxxxxxxxxxxx > Subject: Re: NFS version4 maximum on-the-wire data size. > > > On Sep 1, 2010, at 11:41 AM, Chuck Lever wrote: > > > > > On Sep 1, 2010, at 5:39 AM, Tayade, Nilesh wrote: > > > >> Hi, > >> > >> I am unable to get the exact count for the amount of on-the-wire data > >> size that NFSv4 supports. > >> > >> I could get exact figures in case of NFSv3 > (http://nfs.sourceforge.net). > >> It's 60KB when NFSv3 is used over UDP and implementation dependent in > >> case it is used over TCP. Any idea about NFSv4 statistics? > >> > >> Any pointers would be appreciated. > > > > NFSv4 reads and writes can be as large as the maximum payload size of > the underlying transport. NFSv4 does not support UDP, but on TCP, the > maximum payload size is usually approximately (2 ** 31 - 1) - (RPC header > + NFS overhead). > > But you were probably not asking about the theoretical limits of the > underlying transport. > > The Linux NFS client limits read and write payload size to a megabyte. > Servers may limit it further. Thanks for the information. It helps. Also we use 2.6.22 kernel version (it's old, and have no option of changing it) on our Intel boxes. I was just wondering is there any kernel-version Vs. NFS version mapping (I assume it's not, but just to be sure about)? I have never seen NFSv2 running on our boxes. I am going to write certain applications on top of NFS. So wanted to be sure if NFSv2 should be supported at all (is it so widely used on recent systems, compared to NFSv3/v4)? > > -- > chuck[dot]lever[at]oracle[dot]com > > Thanks, Nilesh nilesh.tayade@xxxxxxxxxxxx -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
