Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 3 Feb 1999 18:16:04 +1030
From:      Greg Lehey <grog@lemis.com>
To:        Robert Clark <Clark@open.org>
Cc:        peter kok <peter@sweda.com.hk>, Greg Black <gjb@comkey.com.au>, FreeBSD Questions <questions@FreeBSD.ORG>
Subject:   Re: Excessive collisions on Ethernet
Message-ID:  <19990203181604.B1179@freebie.lemis.com>
In-Reply-To: <3.0.6.32.19990202225702.0089e840@opengovt.open.org>; from Robert Clark on Tue, Feb 02, 1999 at 10:57:02PM -0800
References:  <36B7CC74.5C09913B@sweda.com.hk> <19990131110224.I8473@freebie.lemis.com> <19990131192548.24006.qmail@alpha.comkey.com.au> <36B7CC74.5C09913B@sweda.com.hk> <19990203144110.W1179@freebie.lemis.com> <3.0.6.32.19990202225702.0089e840@opengovt.open.org>

next in thread | previous in thread | raw e-mail | index | archive | help
[Format recovered at freebie.lemis.com]

On Tuesday,  2 February 1999 at 22:57:02 -0800, Robert Clark wrote:
> At 02:41 PM 2/3/99 +1030, Greg Lehey wrote:
>> On Wednesday,  3 February 1999 at 12:11:32 +0800, peter kok wrote:
>>> Greg Black wrote:
>>>
>>>>> In the last few days I've noticed a really high number of collisions
>>>>> on my Ethernet.  There are only 5 machines on the network, 3 of which
>>>>> are barely active, yet I see:
>>>>>
>>>>> (allegro, running 2.2.6-STABLE)
>>>>> ed0   1500  <Link>      00.00.c0.44.a5.68 43729816    45 43861788 12 977828
>>>>> ed0   1500  widecast      allegro         43729816    45 43861788 12 977828
>>>>
>>>> That's only 1.1%.
>>>
>>> i would like to know how do you measure this collision?
>>
>> This is output from netstat -bi:
>>
>> $ netstat -bi
>> Name  Mtu   Network       Address            Ipkts Ierrs     Ibytes Opkts Oerrs     Obytes  Coll
>> ed2   1500  <Link>      00.80.48.e6.a0.61   358053     5   64720004 458435     0  320307968  8347
>> ed2   1500  widecast      freebie           358053     5   64720004 458435     0  320307968  8347
>>
>> BTW, I had to reboot freebie.  Since then, the incidence of collisions
>> is *much* lower.  There's something funny going on here.
>
> 	I've seen weirdness like this before, but I can't bring the
> cause to the front of my brain at the moment.
> 	I seem to remember one case where a 10-base-T cable had the
> wrong polarity on the rx pair, and this was causing a station to
> miss its collision sense.  Consequently it wasn't doing a good job
> of backing off early.
>
> 	Before anyone gets grumpy with me, I may have the details of
> the cable fault wrong. I also remeber a case where one wire of the
> rx pair was marginal, and the system functioned, albeit
> irrationally.
>
> 	At least once a year I run into someone using two 10-base-T
> nics on the same hub port. How? Some hubs have a port that is both a
> "regular" and "uplink" port. When two station are connected this
> way, the tx collisions slip in under the collison sense
> mechanism. It can be quite a good laugh when you try to explain why
> not to use both ports at once. ( I haven't seen a lan analyzer yet
> that will pinpoint this specific problem. )

Interesting.  That's not the case here, because the whole network is
RG 58.

> 	zzzzzz... oops, I'm curious to see what it turns out to be.

I suspect hardware failure, software bug (it only happens with the ed
driver; other Ethernet boards are unaffected) or phase of the moon.

Greg
--
See complete headers for address, home page and phone numbers
finger grog@lemis.com for PGP public key

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?19990203181604.B1179>