atom feed15 messages in org.freebsd.freebsd-hubsRe: new sup server
FromSent OnAttachments
Paul TrainaMar 21, 1996 11:53 am 
Dima RubanMar 21, 1996 2:52 pm 
Jordan K. HubbardMar 21, 1996 4:04 pm 
Terry LambertMar 21, 1996 5:34 pm 
Paul TrainaMar 22, 1996 9:23 am 
Wayne HernandezMar 22, 1996 10:52 am 
John PolstraMar 22, 1996 1:36 pm 
Peter WemmMar 22, 1996 3:43 pm 
Jordan K. HubbardMar 22, 1996 5:35 pm 
Poul-Henning KampMar 23, 1996 12:00 am 
Michael BeckmannMar 23, 1996 3:13 am 
J WunschMar 23, 1996 2:26 pm 
John PolstraMar 23, 1996 7:08 pm 
Paul TrainaMar 23, 1996 11:32 pm 
Christoph P. KukuliesMar 24, 1996 1:43 am 
Subject:Re: new sup server
From:Christoph P. Kukulies (ku@gilberto.physik.rwth-aachen.de)
Date:Mar 24, 1996 1:43:28 am
List:org.freebsd.freebsd-hubs

From: pet@zit.th-darmstadt.de (Michael Beckmann) Subject: Re: new sup server

I must agree with that. I tried to run a sup mirror in Germany, but have never been able to make world or build a kernel with the sources from that mirror. There has always been corruption of the source tree.

Michael,

as I told you before, I cannot quite share your experience. I supped yesterday -current from sup1.de.freebsd.org to a machine connected locally and the make world ran perfectly without a hitch.

I agree with you that the international lines are stuffed up during daytime in Europe but suping nightly seems to work fine.

I'm surprised. Are you talking -current, -stable, or -cvs?

I build -current trees by suping off of sup2 (i.e. I'm my own customer) every few days and have never had a sup-related corruption.

Due to the highly loaded intercontinental lines, the sup updates from Freefall can take several hours, and in rare cases, (in particular if freefall doesn't let me in due to its ten user limit) they aren't finished until next night. It appears that the situation has improved somewhat with the new Internet connection of Freefall, though, which doesn't route through MCI/BBNPlanet any more.

I would really like to provide some service for up-to-date FreeBSD sources in Europe, but sup simply doesn't work right under these conditions. I wouldn't mind the updates taking several hours, but I do mind getting garbled sources. They cost me a lot of time already.

Sounds like a big call to fix CTM.

--Chris Christoph P. U. Kukulies ku@gil.physik.rwth-aachen.de