Date: Thu, 19 Jul 2007 10:51:11 -0400 (EDT) From: Andrew Gallatin <gallatin@cs.duke.edu> To: =?iso-8859-1?Q?St=E5le?= Kristoffersen <staale@kristoffersen.ws> Cc: current@freebsd.org Subject: Re: Slow networkperformance in current? Message-ID: <18079.31350.366410.316438@grasshopper.cs.duke.edu> In-Reply-To: <20070716233600.GC19282@eschew.pusen.org> References: <20070716190441.GA19282@eschew.pusen.org> <b1fa29170707161247r1a9421cbxf4ec9be1e4c63f7d@mail.gmail.com> <20070716213425.GB19282@eschew.pusen.org> <b1fa29170707161553k356f5ee6vf80dfee7831fc3e0@mail.gmail.com> <20070716233600.GC19282@eschew.pusen.org>
next in thread | previous in thread | raw e-mail | index | archive | help
St=E5le Kristoffersen writes: > On 2007-07-16 at 15:53, Kip Macy wrote: > > Please post the config you are using. >=20 > lrwxr-xr-x 1 root wheel 2B Apr 12 15:42 /etc/malloc.conf -> aj= >=20 > /etc/sysctl.conf: > vfs.vmiodirenable=3D1 > net.inet.tcp.delayed_ack=3D0 > net.inet.ip.forwarding=3D1 > net.inet.tcp.inflight.enable=3D0 >=20 > (I have tried commenting out all those lines, no difference. You definately want to comment out at least net.inet.tcp.delayed_ack=3D= 0 Try temporarily setting net.inet.tcp.sendbuf_auto=3D0 and net.inet.tcp.recvbuf_auto=3D0 and see if that changes anything. I've seen strange issues with IPv6 and the automatic buffer sizing, perhaps you're somehow running into the same thing with IPv4. Drew
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?18079.31350.366410.316438>