Re: [nsp] [long] total packet corruption tricking TCP checksums ?

From: Kai (kai-cisco-nsp-trap@conti.nu)
Date: Mon Jan 31 2000 - 16:40:45 EST


At Monday 02:37 PM 1/31/00 , Alex Bligh wrote:
> > I am having a *big* problem over here, and wonder if any reader of the
> > list has ever experienced something like this (I have a lot of other
> > anecdotal failures going back some 8 years, but never anything like
> > this):
>
>Without wishing to ask the obvious, are you getting any errors
>lik CRC errors *anywhere*? Things being checked by 16 bit CRCs
>will pass 1 in 65536 errored packets (on average).
>
>--
>Alex Bligh
>GX Networks (formerly Xara Networks)

Thanks for pointing out the error probability. I was looking
for this with the help of RFC 1071 just now.

I was getting virtually no CRC errors on the link, but a few giants.
Another reader of this list took the time out to write to me about
Layer 1 and 2 issues in this situation - something that was very
helpful for me (Thanks, you know who you are).

Meanwhile, we've nailed it down to this: whenever one of the 2 UUnet
T1 interfaces is shut down on *UUnet's end only* , things work fine.
As soon as it's configured up (if-no shutdown command), the corruption
occurs. This is independent from my side being shut down, unplugged,
or the circuit put into loopback from one side or the other.

CEF and their HSSI card have become prime suspects this way, I think.

Thanks,
bye,Kai



This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:09 EDT