| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |
To: Derek Atkins <warlord@MIT.EDU>
Cc: Ron Hoffmann <hoffmann@MIT.EDU>, postmaster@MIT.EDU, linux-dev@MIT.EDU
In-Reply-To: Your message of "14 Feb 1999 13:47:08 EST."
<sjmg1886cwj.fsf@datkins.ihtfp.org>
Date: Sun, 14 Feb 1999 15:54:57 EST
From: Greg Hudson <ghudson@MIT.EDU>
Ron wrote:
> The "athena clients" (if we mean cluster machines) are on campus
> and that's different.
Certainly all cluster machines are on campus. There is a smallish set
of vocal people (myself included) who use private Athena machines
off-campus (Grey 17, for example, has several flavors: Solaris,
NetBSD, and Linux). The Athena install doesn't try to distinguish
between on- and off-campus machines.
For the IS-supported platforms and NetBSD-Athena, it would be trivial
to write a "mkserv off-campus-mail" or something which would modify
sendmail.cf to use [outgoing.mit.edu] instead of [ATHENA.$D] as the
relay. Or to queue mail locally, if that's the best answer for
off-campus machines. The Linux-Athena release doesn't use or support
mkserv, but some simple substitute could be devised.
(As an aside, I will note another good reason for cluster workstations
not to queue mail locally: cluster services could reinstall such a
machine at any time, and any queued mail would get lost in that case.
This reasoning applies to a lesser degree to things like lounge
workstations running Linux; the user expectation is that when they
send mail it has gone "off into the mail system" and is no longer
dependent on the machine they sent from. Or so I would think.)
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |