[2206] in linux-net channel archive
Re: Very very bizzare networking problem
daemon@ATHENA.MIT.EDU (Christopher Blizzard)
Thu Mar 21 22:31:07 1996
To: ATPlack@scj.com
cc: alan@cymru.net, linux-net@vger.rutgers.edu, shaver@neon.ingenia.com
In-reply-to: Your message of "20 Mar 1996 09:31:42 CST."
<"MSMAIL PC */PRMD=SCJMAIL/ADMD=MCI/C=US/"@MHS>
Date: Thu, 21 Mar 1996 21:32:49 -0500
From: Christopher Blizzard <blizzard@odin.nyser.net>
In message <"MSMAIL PC */PRMD=SCJMAIL/ADMD=MCI/C=US/"@MHS>, ATPlack@scj.com wri
tes:
>
>Two identical IP addresses on the Physical segment that you are having
>problems with. If you can RARP the address with the Linux box off. This
>will give you the MAC of the board causing the problem.
>
>The connection was refused due to the other IP station replying to your
>address.
>
Well, looks like I get to bite the bullet on this one. There was
a Mac on my subnet with the same IP address. This is why I missed it on
the first try. It's not on all the time. :) I still wonder why tcpdump
showed the same MAC address though. Odd. Thanks for all your time,
guys. Appreciated.
--Chris
ps. Wow. Do I feel like an idiot. :)
>The 1.3.57 code is helping because of the new way that the NET code is
>written.
>
>Good luck.
> ----------
>From: blizzard@odin.nyser.net
>To: alan@cymru.net
>Cc: linux-net@vger.rutgers.edu
>Subject: Very very bizzare networking problems. :
>Date: Wednesday, March 20, 1996 12:29AM
>
>Precedence: bulk
>
>
> Howdy. I have been experiencing a very odd network related
>problem and was hoping you might have the time to help. I'm also going to
>cc: this to the linux-net mailing list in hopes that someone might have
>something to offer there as well. :) Here is the description:
>
> I originally ran 1.2.13 when I first set up this Linux box.
>Unfortunately, I found that if I didn't generate network traffic every so
>often, say every 15 minutes or so, that I would no longer be able to
>access the machine remotely (An ARP problem of some kind?). Any service
>that I tried to use be it web services, ftp, telnet, or sendmail would
>respond with "Connection refused." If I went to the console and telnetted
>out I could then telnet back in without problems. There were a couple of
>times when I pinged the machine from offsite and it would return two echos
>instead of one.
>
> I figured that there might be something wrong with the kernel, so
>I thought I'd upgrade. I tried running 1.3.57, a damn stable kernel, and
>see if it would help. It did seem to. I can now telnet in without
>problems as long as I am not on the same physical cable as the machine.
>If I am the same segment, I still receive the same messages as above.
>However, if I try again and again (especially from one of the suns on the
>segment) to connect to port 80 I will eventually get a connection. At
>that point I can connect without problems from any machine on that
>segment to that service.
>
> A co-worker of mine also runs Linux. He does not seem to
>experience the problems that I do. We run on the exact same hardware.
>The only difference that I can figure is that he has 10Base-T, and I use
>the AUI port with an adapter to Coax on my ether card. We both have
>3Com 3C579's. I've tried booting using his kernels off of disk and I
>have exactly the same problems. Any thoughts on how I might diagnose
>this further? Thanks.
>
> --Chris
>
> ---------------------------------------------------------------------
>Christopher Blizzard | "The truth knocks on the door and you say
>blizzard@odin.nyser.net | 'Go away. I'm looking for the truth,' and
> | so it goes away." --Robert Pirsig
> ---------------------------------------------------------------------
>
>
>
