NFS performance tuning: a benchmark and a question - NFS

This is a discussion on NFS performance tuning: a benchmark and a question - NFS ; OK, I know NFS isn't usually thought of as the fastest protocol under the sun, but still, there are times when making NFS move along a little faster can be worthwhile. I've written a sort of NFS benchmark that I'm ...

+ Reply to Thread
Results 1 to 4 of 4

Thread: NFS performance tuning: a benchmark and a question

  1. NFS performance tuning: a benchmark and a question


    OK, I know NFS isn't usually thought of as the fastest protocol under the
    sun, but still, there are times when making NFS move along a little faster
    can be worthwhile.

    I've written a sort of NFS benchmark that I'm calling nfs-test. It
    tries a largish number of rsize's, wsize's, tcp vs udp, and version 2
    or 3 (4 would be very easy to add), to see what gives the best
    performance. You can find it at
    http://dcs.nac.uci.edu/~strombrg/nfs-test.html

    My question is, before diving into trying to determine this empirically,
    is there any theoretical reason why it would be better to have
    rsize==wsize, or should it be better to just pick whatever rsize gives the
    best read performance and pick whatever wsize gives the best write
    performance, and not worry about if rsize!=wsize?

    Thanks!


  2. Re: NFS performance tuning: a benchmark and a question


    "Dan Stromberg" wrote in message
    newsan.2005.09.06.18.12.58.160981@dcs.nac.uci.edu...
    >
    > OK, I know NFS isn't usually thought of as the fastest protocol under the
    > sun, but still, there are times when making NFS move along a little faster
    > can be worthwhile.
    >
    > I've written a sort of NFS benchmark that I'm calling nfs-test. It
    > tries a largish number of rsize's, wsize's, tcp vs udp, and version 2
    > or 3 (4 would be very easy to add), to see what gives the best
    > performance. You can find it at
    > http://dcs.nac.uci.edu/~strombrg/nfs-test.html
    >
    > My question is, before diving into trying to determine this empirically,
    > is there any theoretical reason why it would be better to have
    > rsize==wsize, or should it be better to just pick whatever rsize gives the
    > best read performance and pick whatever wsize gives the best write
    > performance, and not worry about if rsize!=wsize?
    >
    > Thanks!
    >


    Dan,

    Interesting... Did you notice
    http://nfs.sourceforge.net/nfs-howto/performance.html
    has a section on benchmarking and tuning that covers
    pretty much the same thing that you just did ? Hmmm....

    Enjoy
    Postmaster



  3. Re: NFS performance tuning: a benchmark and a question

    On Tue, 06 Sep 2005 19:28:37 +0000, Postmaster wrote:

    >
    > "Dan Stromberg" wrote in message
    > newsan.2005.09.06.18.12.58.160981@dcs.nac.uci.edu...
    >> [quoted text muted]

    >
    > Dan,
    >
    > Interesting... Did you notice
    > http://nfs.sourceforge.net/nfs-howto/performance.html
    > has a section on benchmarking and tuning that covers
    > pretty much the same thing that you just did ? Hmmm....
    >
    > Enjoy
    > Postmaster


    Thanks for pointing this out. I hadn't read it. I was aware of bonnie,
    iozone, dbench and such though.

    The script I mentioned/wrote probably could be used for any of these (but
    currently uses my reblock program) - to automate trying a bunch of rsize's
    and wsize's. I may be wrong, but I currently believe these programs are
    for testing different block sizes, rather than rsize's and wsize's and
    tcp/udp.


  4. Re: NFS performance tuning: a benchmark and a question

    The "block size" mentioned in the link is the max rsize/wsize that the
    server will support.


+ Reply to Thread