Date: Sun, 16 Jan 2000 11:23:22 -0500 From: "C J Michaels" <cjm2@earthling.net> To: "Andrew Reilly" <areilly@nsw.bigpond.net.au> Cc: "FreeBSD Questions" <freebsd-questions@FreeBSD.org> Subject: RE: (fast) ethernet performance problems/tweaking Message-ID: <NDBBJKPOALBHJNGOLOFNAEBBCAAA.cjm2@earthling.net> In-Reply-To: <20000110075748.A29687@gurney.reilly.home>
next in thread | previous in thread | raw e-mail | index | archive | help
Well, all I can say is, do you have a switch or a hub? If you have a hub,
odds are that it doesn't support full-duplex in the 1st place and that's why
you are getting really poor performance.
Try forcing the cards to half-duplex and see what happens.
-Chris
-----Original Message-----
From: owner-freebsd-questions@FreeBSD.ORG
[mailto:owner-freebsd-questions@FreeBSD.ORG]On Behalf Of Andrew Reilly
Sent: Sunday, January 09, 2000 3:58 PM
To: questions@FreeBSD.ORG
Subject: (fast) ethernet performance problems/tweaking
Hi,
I have a little network at home consisting of my
FreeBSD-3.4-STABLE box (a PIII-500) and a Windows-NT 4.0 box (a
Celeron-400). Since there are only two machines, I use a
crossover cable instead of a hub or switch.
In a recent fit of upgrading, I replaced the 10-baseT (PCI) cards in
each machine with a pair of 100-baseTX RealTek-8139 cards. I
was pleased that everything just seemed to work, but I've just
tried to test the performance, and to say that it's short of
stellar is an understatement.
I have a 16M file in my home directory (FreeBSD), and two
successive command-line FTP fetches on the NT box resulted in
transfer rates of 70.02k and 99k. Yes, "k". I tried using the
"copy /b" command, and gave up timing after five minutes.
Where can I look to try to debug what is obviously a problem?
Here's a bit of representative output from a
netstat -I rl0 -b -w 5
command, while the copy/b was in progress:
input (rl0) output
packets errs bytes packets errs bytes colls
42 0 2636 81 0 118620 0
41 0 2692 71 0 100806 0
37 0 2336 72 0 104994 0
30 0 1916 57 0 80944 0
32 0 2152 57 0 80950 0
58 0 3712 97 0 141510 0
40 0 2516 72 0 104994 0
35 0 2216 64 0 92882 0
Here's the output of
ifconfig rl0:
rl0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
inet 10.0.0.1 netmask 0xff000000 broadcast 10.255.255.255
ether 00:48:54:50:52:83
media: 100baseTX <full-duplex>
supported media: autoselect 100baseTX <full-duplex> 100baseTX
<half-duplex> 100baseTX 10baseT/UTP <full-duplex> 10baseT/UTP 10baseT/UTP
<half-duplex>
Hmm That's interesting. Shouldn't the flags say DUPLEX instead
of SIMPLEX when the media is in full-duplex mode?
Here's some /var/log/messages output related to rl0:
Jan 6 13:35:51 gurney /kernel: rl0: <RealTek 8139 10/100BaseTX> rev 0x10
int a irq 9 on pci0.10.0
Jan 6 13:35:51 gurney /kernel: rl0: Ethernet address: 00:48:54:50:52:83
Jan 6 13:35:51 gurney /kernel: rl0: autoneg not complete, no carrier
Jan 6 13:35:51 gurney /kernel: rl0: selecting MII, 100Mbps, half duplex
Jan 6 13:35:51 gurney /kernel: rl0: selecting MII, 100Mbps, full duplex
in rc.conf I have:
ifconfig_rl0="inet 10.0.0.1 media 100baseTX mediaopt full-duplex"
Unfortunately I know even less about NT networking than I do
Unix networking, so I don't know where to start, for checking
the NT end of the link.
Thanks in advance for any suggestions,
--
Andrew
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?NDBBJKPOALBHJNGOLOFNAEBBCAAA.cjm2>