Date: Wed, 1 Feb 1995 09:06:10 -0500 (EST) From: "Michael C. Newell" <mnewell@lupine.nsi.nasa.gov> To: Michael Nelson <nelson@seahunt.imat.com> Cc: James Robinson <james@hermes.cybernetics.net>, questions@FreeBSD.org Subject: Re: Discouraged... Message-ID: <Pine.SUN.3.91.950201085323.5821F-100000@lupine.nsi.nasa.gov> In-Reply-To: <199501312115.NAA01599@seahunt.imat.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 31 Jan 1995, Michael Nelson wrote: > > > What speed -- 16 cps ? Are you saying that you get 1.5K/s from your > > FreeBSD machine to other machines in the real world, but really > > slow ones to freefall.cdrom.com? > > Yes, that's exactly what I'm saying. For example, I tried to sup > again today. It took an HOUR to get these three files: > > -rw-r--r-- 1 root wheel 3425 Sep 11 00:53 COPYRIGHT > -rw-r--r-- 1 root bin 9729 Jan 24 19:00 Makefile > -rw-r--r-- 1 root bin 5843 Jan 29 01:00 TODO > > ...then I aborted it again. > > > Or do you get 1.5K/s from a peer Linux machine to the real world? > > That's also true. > > > Can you capture some documentary output describing your situation, as > > well as your network setup? > > Sure... but what information do you want to see? Traceroutes? > Pings? Tcpdumps? Let me know and I'll provide them. > > > Ok -- what is acting as the gateway machine? > > A Compaq i386/20DX with 13MB of RAM, running Linux and Dip. It's > a dedicated slip machine with nothing else going on on it. I ran into the identical problem using a Lunix machine as a gateway; except that my end-host was a Sun IPX running SunOS 4.1.3. FTPs took forever, and when completed often the files were shorter than the originals! We also couldn't get X to run through the Linux box. So... ... we upgraded the Linux kernel to the latest version (from TSX-11). All our problems went away. It appears that the networking code was badly broken and the new kernel fixed it (we were using the Slackware V2.0 distribution as I recall, from the Linux Developer's Kit; this was about 6 months ago?) What kernel version are you running on the Linux box? We're running 1.1.18 on our gateway machine now with no problems. [related topic] One thing that helped dramatically improve the performance of my FreeBSD gateway [we're playing with lots of machines here running lots or OSen to see how different systems do as gateways/routers] was to up the SLIP MTU from 296 to 576 (?). I almost doubled my throughput for FTP transfers, and still don't have horrible response when FTPing and TELNETting at the same time. The 296 MTU was fine in the old days of 2400 baud modems; it don't cut it on 28.8!! [end related topic] > > > What does the FreeBSD box's routing table look like? > > I don't know what I need to do to get you that information. Can > you tell me what you need for me to do? I plead "dummy" on this one. > You do netstat -nr The "n" tells it not to translate addresses to names; I find it more useful. Also you should do a netstat -ni to get a dump of the interface configurations. > As I said above, it _is_ a DX chip. Wasn't there a problem reported some time back with AMD 486 chips? I don't have one so I didn't pay much attention, but I seem to remember a thread about problems caused by a bug in the chip itself [no, I'm NOT confusing it with the dePentia chip!] Anyway, might be worth a look through the Email archives... Thanks, Mike +--------------------------------------+------------------------------------+ |Mike Newell | The opinions expressed herein are | |NASA Science Internet Network Systems | my own, and do not necessarily | |Sterling Software, Inc. | reflect those of the NSI program, | |MNewell@nsipo.nasa.gov | Sterling Software, NASA, or anyone | |+1-202-434-8954 | else. | +--------------------------------------+------------------------------------+
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.SUN.3.91.950201085323.5821F-100000>