Re: Possible TCP Problem with RH6.2 talking to Solaris2.6/2.7

From: shane (shane@bratnet.net)
Date: Thu May 04 2000 - 14:25:48 EST


Just to catch you up, that dump was created from the linux side,
monitoring the ethernet segment that was connected to the solaris box
with a flip cable and the sparc is locked at 100BaseTX-FD and mii-diag
showed that the Intel interface that was on the otherside was also
100BaseTX-FD.. I tried switching 1 side to HD and the perf went down to
50-500kB/sec eek.

On Thu, 4 May 2000 kuznet@ms2.inr.ac.ru wrote:

> Hello!
>
> > More Info on this is available by emailing me, unless this problem can get
> > fixed, I will have to pull the linux installed servers and replace with
> > FreeBSD 4.0,
>
> Is it something sort of blackmailing? It is interesting. 8)
> OK, OK, no problems. 8)
>
> Question #1 (and last for beginning): kernel version?
> I cannot identify 172.16.10.2 as any known linux kernel.
> It behaves differently both of 2.2 and 2.3. It is critical moment.
>
> Analysis made by that expert is correct, _assuming_ that
> dump is correct. But it is not. It is pretty evident, that
> this dump does not show lots of packets.
>
>
> > Well I have tried EVERYTHING, originaly it was a ful-duplex problem,
> > and I fixed that with the mii-diag program.
>
> Go to switch and make sure that _both_ ends think that
> they are full-duplex. Packet lossage on 100Mbit links are
> almost always explained by failed full-duplex negotiation.
>
> Also, if port going to one machine is full-duplex, and
> other one is half-duplex, most of switches drop lots
> of packets at peak loads.
>
>
> To resume: it looks like you have massive lossage in ACK path
> i.e. in direction linux -> solaris. This can be explained
> only by wrong switch setup.
>
> Alexey
>

-
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 - 21:00:15 EST