Date: Thu, 12 Nov 1998 11:11:47 -0500 From: "Jason J. Horton" <jason@intercom.com> To: current@FreeBSD.ORG Subject: Re: Is it soup yet? FreeBSD NFS Message-ID: <364B08C3.27FE59A4@intercom.com> References: <Pine.BSF.4.05.9811112154240.370-100000@porkfriedrice.ny.genx.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Everyone seems to be talking about using FreeBSD as an NFS client,
how does FreeBSD do as a NFS server?
-J
Alfred Perlstein wrote:
> On Wed, 11 Nov 1998, Brian Feldman wrote:
>
> > > > When we "catch up" to Linux, for every advance, we always have a
> > > > better-implemented version of whatever new has been gotten on Linux. Maybe
> > > > except for NFS.... but that's being working on, eh?
> > >
> > > I hope you are refering to Linux NFS being far inferior to FreeBSD's
> > > impelementation. If you are not then where do we fall behind? I haven't
> > > seen Linux outperform FreeBSD in any NFS work i've done.
> >
> > Last I had heard, NFS was still too unstable to be used heavily in
> > FreeBSD, but "worked" in Linux. But, who knows for sure? I don't run any
> > Linux systems, and I don't really use NFS in FreeBSD.
>
> Then perhaps you should stay quiet on the issue.
>
> FreeBSD has outperformed linux by several orders of magnitude in client
> side NFS for a long time. The newer linux development kernels come close,
> but when concurrent NFS requests are made Linux chokes while FreeBSD
> maintains a broadband'ish state.
>
> I know i posted several times about problems with NFS (about a month ago)
> but since McKusik's fixes I've yet to have a problem.
>
> In so far as serving NFS... the Linux userland NFS server is hardly a
> match to the FreeBSD kernel impelementation. Stability is another matter
> and I haven't seen enough to say anything conclusive for either side.
> Both implementations suffer from lack of support for files > 2gb in
> client side requests which should be addressed, somehow/somewhen.
>
> A recent Linux article suggests that Linux NFS will bipass the "mbuf"
> layer, ie. the NFS code will directly reassemble packets into RPC requests
> thereby saving _one_ copy of memory. This is really neat, but then makes
> NFS dependant on the protocols which it is supposed to be independant of.
>
> Btw, Mike Smith's new ACCESS caching seems quite stable and i was
> wondering if it had been commited.
>
> -Alfred
>
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-current" in the body of the message
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?364B08C3.27FE59A4>