On Mon, 15 May 2000, Michael Kukat wrote:

[...]
> So i think, excessive UDPing might make trouble. I didn't try
> excessive ICMPing by doing a ping -f, this might be another nice
> test. 

I tried flood ping for 10 minutes. Some 22 packets (out of 400.000)
dropped. 37.6 MB ICMP traffic according ntop. No xmit error...

And then a very short telnet connection and yes, there it is
again. (After closing the telnet connection. Funny.)
Any longer telnet session now ends here with a reboot though.

Just tried to work solely on the console, but same xmit stuff. It
keeps coming all the time now ad nauseam. 

It very much reminds me of problems I had with a diskless sun 3/80 I
had for some time. I cannot remember the precise messages, but it was
some timeout of the le0. Only a hardware reboot (I mean turning the
box off and on) was helpful IIRC.

The ze0 problems are with the GENERIC kernel of May 13 sources. (I
wanted to make sure my custom kernel was not the cause. I had no BOOTP
in it or INET6 and some devices my machine hasn't. Couldn't think how
this could matter, but that proves nothing.)

I am very new to the vax, but I have the impression this problem is
much bigger in the recent snapshot (from neurotica, source of 24
April) than in the earlier one (november) I had for a few days. Could
I be right?

> In pure TCP, the SGEC works wonderful, and has a nice
> performance. Maybe i should try mounting the NFS stuff with TCP,
> this could improve stability.

How? Just tell nfsd on the server to accept only TCP does *not* work.
(Booting stops at @@@@@@@@@/). There is nothing in man nfsiod about
it. /etc/services?

--
Marc Schneiders --- marc@venster.nl --- marc@vrijdom.com

FreeBSD unclad.freebeastie.org 5.0-CURRENT