On Sun, 7 May 2000 kuznet@ms2.inr.ac.ru wrote:
> Hello!
>
> > eth0 1500 0 65251461 34 0 12 36945441 0 0
> > 0 BRU
>
> Wow! It overruns... No, it is not bus speed. I do not know what it is.
> I never saw overruns due to tcp in my life.
My machine is showing 3 TX overruns on eth0, having RXed 17.6m packets and
TXed 600k packets. Probably not TCP related, though. (And not enough of
them to worry about, I suspect.)
> > Well my guess is that while the Solaris box is under load, and the linux
> > box is under load, It causes the Case 1, in the original posting to occur?
>
> Ough. Please, please, make tcpdump.
That should help shed some light on the situation...
> > Case 1)
>
> Seems, I have to repeat, this one was not tcpdump from linux. 8)8)
This doesn't help :)
> > Would you belive that the lack of multiple ACK's is a Solaris problem,
> > that sould be reported as a Solaris Bug, or something that like sujected
> > is a error on the linux side?
>
> If that tcpdump is really tcpdump of solaris receiver, it definitely
> shows bugs. It really does not send duplicate acks and sender
> cannot recover from single loss.
>
> But real root of problem is not here. If ethernet is so fragile,
> you will have lots of troubles even if TCP stack is flawless.
> At least, it would not not fall to timeout in this case,
> if receiver were linux.
There is definitely something amiss...
Perhaps you could try that 3com NIC instead of the EtherEpress 100? **and
make sure the machines are NOT loaded**.
James.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun May 07 2000 - 21:00:21 EST