[547] in Info-AFS_Redistribution
Re: Changing DB server IP addresses
daemon@ATHENA.MIT.EDU (Daniel Edward Lovinger)
Mon Jan 20 13:47:47 1992
Date: Mon, 20 Jan 1992 12:44:28 -0500 (EST)
From: Daniel Edward Lovinger <dl2n+@andrew.cmu.edu>
To: Info-AFS@transarc.com
In-Reply-To: <Added.YdSfF7=0Bi81A43k9A@transarc.com>
"Peter Lister, Cranfield Computer Centre" <CC012@cranfield.ac.uk> writes:
> > I believe the most correct solution is to never change the IP
> > addresses of your DB servers once you've advertised their existence.
> > That's the approach we take here.
>
> This may be a useful workaround, but does not represent a "correct
> solution".
Agreed. CMU's network configuration is a rare case in the real
world and to the extent that that affected the development of some the
early AFS innards it is probably unfortunate.
> I'd like to add my voice to the request for a properly delegated
> distributed database, BIND/Hesiod style.
From that which I have heard of DFS, you get this in OSF/1. By
appropriate magic, paths of the form
/.../edu/cmu/andrew/fs
(or is that /.../andrew.cmu.edu/fs ?)
will be the normal way to get to a cell. I could think of more
important things for Transarc to do to AFS 3, though. Rem: afs 3 is
moving toward maintainence mode.
> Until then, Transarc should
> supply a program which regularly reports back local server addresses,
> and picks up the latest authoritative CellServDB, so that we don't all
> have to do it manually.
It would not be difficult to construct a cron-ish facility to
do this if transarc.com is mounted and you have your machines running
at least a dwarf package configuration ...
D /
FA /usr/vice/etc/CellServDB /afs/transarc.com/service/CellServDB root wheel 644
(and run every week ... reboot "often")
> It should not rely on the cell having
> /afs/transarc.com mounted, because there may be good reasons not to.
Why? Do you worry about 'ls -F' on /afs? It isn't a
requirement to mount cells there (some sites mount under /afs/external
or similar, though this kinda defeats the "global namespace"). I think
it is reasonable for Transarc to expect customers to be able to pick
things out of their cell (and it's faster than ftp).
Dan Lovinger Computing & Communications Carnegie Mellon U.
Internet: dl2n+@andrew.cmu.edu Bitnet: dl2n+%andrew.cmu.edu@carnegie
"... and in the stillness, they heard the cry of the golden banana ..."
