Date: Thu, 1 Dec 2005 21:09:06 +0000 From: Freminlins <freminlins@gmail.com> To: "N.J. Thomas" <njt@ayvali.org> Cc: freebsd-questions@freebsd.org Subject: Re: overloaded webserver: nfs wait issue? Message-ID: <eeef1a4c0512011309ke3feff0rb609131dd2dd0742@mail.gmail.com> In-Reply-To: <20051201193813.GG15171@ayvali.org> References: <20051201193813.GG15171@ayvali.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Can you provide the output of nfsstat -c 1 on the web servers (and nfsstat -s 1 on the nfs server if they are running FreeBSD)? Run them for a minute when you are getting bad performance. Please also quantify "moderately high traffic". What is this in Mbit/s? Also, it may be helpful to know a little bit about your network infrastructure, i.e. are the switch ports set to full duplex and the machines to match. NFS performance is very dependent on a fast reliable network. Don't consider anything less than 100Mbit/s full duplex. Also what platform is the NFS server? We recently migrated from a cluster pair of NetApps to a single NFS server running FreeBSD. This is mounted by four NFS clients running Zeus. We have = a lot of NFS experience and we felt confident this would work. It did and wit= h the most marginal degradation in performance. Frem.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?eeef1a4c0512011309ke3feff0rb609131dd2dd0742>