[ntp:hackers] Re: Another Broadcast client fix

David L. Mills mills at udel.edu
Thu Jan 8 12:24:00 PST 2004


Danny,

No big deal at all either way, just that we don't get long out of sync
with ntp-dev. It is true I sometimes have short-term experiments going
on some machines, but none except malarky and mort are tied up with
long-term ones.

We have had several flaky-net incidents recently, including a 128-4
black hole during the Christmas break, a department Ethernet flake
yesterday and the usual intermittent power outages and ISDN spasms. The
staff are warming up a new Sun for me to replace power-flake mort; the
other machines are doing fine.

Happily, I am told the new OC-3 Delaware Ring is now operational, which
replaces flaky 10-Mbps via ISP Voicenet, so our commercial interconnect
is now almost as good as our Internet-2 interconnect. The staff is in
progress of upgrading our wires and routers to gigabit Ethernet and the
campus is already there. Mike thinks the average aggregate rate past our
128.4 routers is much less than a megabit. Kinda like cosmic background
radiation with occasional supernova explosions. Nice to have that
horsepower.

Dave

Danny Mayer wrote:
> 
> At 09:56 AM 1/7/04, David L. Mills wrote:
> >Danny,
> >
> >Using campus test machines is a little dangerous unless you verify first
> >that the ntp-dev stuff on whimsy/mort has been copied to pogo and that
> >any changes on pogo be retrofitted to whimsy/mort. Nevertheless,
> >bridgeport seems a success.
> 
> Dave,
>          Normally I wouldn't, but I had less than an hour to figure this out
> and all of my bits are on pogo. bridgeport didn't seem to be doing much
> so I just went for it. Sorry. Dealing with the backroom is much harder
> and I don't know what experiments you may be running on them.
> 
> Danny



More information about the hackers mailing list