Chris Galle <[email protected]> wrote:
> Not that temperature is not the issue, but I will point out that everyone
> that is not having problems is running *BSD. Could this be a driver issue
> with the particular version of the linux kernel you are running?
I've seen problems on non-soekris machines that result not in a
permanent freezeup or kernel panic but instead kill the TCP connection
part way through file transfers (mostly large, but sometimes small).
Decide for yourself if this is a red-herring in this case.
The problems exhibit themselves with some physical NICs in non-VM
environments, and in VM environments (specifically VMWare and Xen,
but not necessarily limited to those environments). It comes down to
enabling TCP segment offload, but either the physical NIC or the
virtual stack not doing it correctly, thus ending up with corruption
of the TCP session. You can see if it's enabled by running:
ethtool -k eth0
and you can disable it via:
ethtool -K eth0 tso off
I don't know if this problem exhibits itself on the Soekris machines
as I've never installed Linux on a Soekris; I use OpenBSD there.
Devin
--
I got food poisoning today. I don't know when I'll use it.
- Stephen Wright
_______________________________________________
Soekris-tech mailing list
[email protected]
http://lists.soekris.com/mailman/listinfo/soekris-tech
