After they came back up, there was a tremendous volume of extraneous
network traffic. For about 15 minutes, the Server's CPU meter was
pegged. I would not be surprised if the server was dropping a great
many packets during this 15 minute interval. Obviously all packets
were being dropped for the preceding 50 minutes.
Things seem to be back to normal. The server is running at about
50% CPU utilization, and it's log is again scrolling in sync with
the HUB's activity light.
Dan's previous advice was good -- DO NOT PANIC. I think a part
of the extraneous network traffic were traceroutes and pings.
Just so you folks know... There is a shiny new Pentium 133 sitting
here, ready to take the place of the crusty old 486DX2-66 keyserver.
The Pentium would have cleared that 15 minute surge somewhat more
quickly, but otherwise really wouldn't have had any measurable effect
on our overall keyspace rate.
Thanks for your patience!
> From email@example.com Fri May 9 20:26:40 1997
> Date: Fri, 9 May 1997 20:22:21 -0600
> To: firstname.lastname@example.org
> From: email@example.com (Dan Oetting)
> Subject: Re: DO NOT PANIC -- the server is not down
> It looks like the server is responding normally again. For a brief time
> this evening it was only responding to about 1 in 30 packets. This may have
> been caused by network congestion at frii.com or an overload of the server
> itself. I can only speculate based on the packet traffic at my site because
> I have no official standing in the deschall project.
> -- Dan Oetting <firstname.lastname@example.org>