Linux 2.0.27, P100, ISA SMC cards, 10baseT network and this box is connected
as the only link into a UB GeoRim (Switch) although the same occurs when it is
on a network with other PC's.
> Now, what is happening is this: the throughput is extremely "sluggish". If
> I look at the 10BaseT Ethernet hub, what I observe is this:
>
> <fast blinking of LEDs for both machines, some collisions> - ~ 1 sec.
> <PAUSE - all traffic LEDs dead> - ~ 3 sec.
> <fast blinking ... > - about 1 Seconds
> <PAUSE> - about 3 Seconds
>
> *and so on*
> The rhythm is *very* regular, both "watching the net" and on the
> performance meter.
> Now... has anybody of you *any* idea what is happening here? I've searched
> the Kernel sources for something like "we have seen too many collisions,
> let's back off and try again in a couple of seconds" or so, but couldn't
> find anything.
> gert
> --
> email: gert@greenie.muc.de fax: +49-89-3244814 http://www.leo.org/~doering
Mine is a bit different and I mentioned it earlier so I hope noone objects if
I recap :):
I have a situation where the line seems dead. PC's can ping other boxes, but
between the 2 boxes communicating they seem to block. It appears as a hang on
X servers waiting for data. After a time period from a second to a minute or
worse, they take off and catch up and are ready to go again. A good program to
make it happen is a graphic intensive, real-time VMS app that just _hangs_
once it starts.
We put a fluke meter on the line and it showed an unaccpetable level of
collisions on the line. This made us put it on a separate segment with a
direct link to our switch. Still, by viewing the switch statistics, we are
seeing way too many collisions.
Another oddity is that in a repeated test, sending data out from the Linux box
to the switch caused next to no collisions. Pulling in a 2meg file to the
Linux box caused hundreds of collisions. 1000 packets, 300 with colisions, not
too good :).
We have tried changing cards, and changing cabling with no effect. I am next
going to try a PCI card, but have no expectations :(.
So maybe this is completely different, but then again ;),
JES
-- James B. MacLean Department of Education and Culture macleajb@ednet.ns.ca